Stellar Death in the Nearby Universe

DISSERTATION

Presented in Partial Fulﬁllment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Thomas Warren-Son Holoien

Graduate Program in Astronomy

The Ohio State University 2017

Dissertation Committee: Professor Krzysztof Z. Stanek, Advisor Professor Christopher S. Kochanek, Co-Advisor Professor Todd A. Thompson Copyright by

Thomas Warren-Son Holoien

2017 Abstract

The night sky is replete with transient and variable events that help shape our universe. The violent, explosive deaths of stars represent some of the most energetic of these events, as a single star is able to outshine billions during its ﬁnal moments.

Aside from imparting significant energy into their host environments, stellar deaths are also responsible for seeding heavy elements into the universe, regulating star formation in their host galaxies, and affecting the evolution of supermassive black holes at the centers of their host galaxies. The large amount of energy output during these events allows them to be seen from billions of lightyears away, making them useful observational probes of physical processes important to many fields of astronomy.

In this dissertation I present a series of observational studies of two classes of transients associated with the deaths of stars in the nearby universe: tidal disruption events (TDEs) and supernovae (SNe). Discovered by the All-Sky Automated Survey for Supernovae (ASAS-SN), the objects I discuss were all bright and nearby, and were subject to extensive follow-up observational campaigns. In the ﬁrst three studies, I present observational data and theoretical models of ASASSN-14ae,

ASASSN-14li, and ASASSN-15oi, three TDEs discovered by ASAS-SN and three

ii of the most well-studied TDEs ever discovered. Next I present the discovery of

ASASSN-13co, an SN that does not conform to the traditional model of Type II

SNe. Finally, I discuss the full sample of bright SNe discovered from 2014 May 1 through 2016 December 31, which is signiﬁcantly less biased than previous nearby

SN samples due to the ASAS-SN survey approach, and perform statistical analyses on this population that will be used for future studies of nearby SNe and their hosts.

iii Dedication

To my grandfather, Warren Dean Chinn, who would have loved to read this.

iv Acknowledgments

My unorthodox path to writing this dissertation would not have been possible without the support, kindness, friendship, and guidance of many individuals. While

I am only able to name a few of them here, I am grateful to all the others who have been instrumental to my success throughout my life.

I would ﬁrst like to thank my Ph.D. advisor, Kris Stanek, for his lessons about life and science during the course of my graduate studies. Kris taught me that when it comes to science, “keep it simple, stupid,” and “perfect is the enemy of the good” are invaluable words to live by, and I would not have achieved nearly the same level of success without his guidance and encouragement. Kris was always available but never overbearing, was extremely accommodating to my living situation, and always had a joke ready to lighten the mood when I was bogged down by research. It has been a privilege to work with Kris, and I am proud to be one of his students.

I am thankful to Chris Kochanek for being available and able to answer any questions I had about science, ASAS-SN, or pretty much anything else during my four years at Ohio State. Chris never failed to provide feedback on paper drafts within 24 hours of receiving them, and he has helped me improve my academic

v writing more than anyone. If I can emulate even a fraction of his dedication and discipline in my own career, I know I will be successful.

I am grateful to Todd Thompson for showing me how to approach learning new material in a way that allows me to more efficiently remember the most important facts. Todd’s ability to discuss any field of astronomy at an order-of-magnitude level is truly impressive, and his example has inspired me to improve the breadth of my knowledge. Todd was also an extremely effective spokesperson for the department in the year I applied to Ohio State, and I would not be here without his influence.

I thank my undergraduate advisor, Saurabh Jha, for giving me my ﬁrst research experience and for encouraging me to apply to OSU for graduate school. I also thank

Risa Weschler and Phil Marshall for advising me during my summer at SLAC.

Ben Shappee has been both a good friend and a mentor to me since I came to

Ohio State, and I appreciate his quick acceptance of my joining the ASAS-SN team.

I hope I can continue to (literally) follow in his career footsteps in the future. Jose

Prieto has also been an important mentor to me, and I am grateful for his help with data analysis and his guidance on navigating my graduate career. Jon Brown has been both a great friend at Ohio State and my oﬃce mate for three years, and I thank him for providing insightful conversation, lightening the mood, and helping to handle the day-to-day work of ASAS-SN. I am also grateful to the many others who work on ASAS-SN, without whom the survey would not be nearly as successful.

vi I thank the entire OSU Department of Astronomy for fostering an environment where science is held in high regard, collaboration is encouraged, and graduate students are valued both for their opinions and as people. I am thankful to the OSU

Graduate School, the Center for Cosmology and AstroParticle Physics, and the US

Department of Energy for ﬁnancial support over the past four years.

I would like to close by thanking those whose support and love have impacted my life the most: my family. Mom, thank you for being there for me no matter what, and for always pushing me to do my best. Dad, you showed me how amazing this Universe is when we looked at the skies together. Thank you for inspiring me and showing me how to enjoy life. Peter, thank you for providing me with laughter, friendship, and countless nicknames over the last 26 years. Melani, thank you for being the big sister I never had. Nana and Grandad, you provided me with the

ﬁnancial freedom to pursue a career in astronomy and you have been my most vocal cheerleaders throughout my life. I cannot thank you enough. Son family, thank you for accepting me into your family and supporting me like one of your own. To my extended family: I love you all and could not have asked for better relatives. Most importantly, Deborah: you are the most loving, supportive, and kind person I know, and I never would have reached this point without you. You have made me a better scientist and a better person, and I will always be striving to be worthy of you.

Finally, I thank God, without whom the wonders of this Universe would not exist, for guiding me down the long and winding path that led me here.

vii Vita

June 2, 1986 ...... Born – Stanford, CA, USA

B.A., East Asian Studies, Stanford University, 2008...... Stanford, CA, USA

SAS Excellence Award, Class of 1925 Scholarship, 2013 ...... Rutgers University, New Brunswick, NJ, USA

B.S., Astrophysics, Rutgers University, 2013 ...... New Brunswick, NJ, USA

Center for Cosmology and AstroParticle Physics 2013 ...... Early Start Award, Columbus, OH, USA

University Fellow for Graduate Studies, 2013 ...... The Ohio State University, Columbus, OH, USA

US Department of Energy 2014 – 2017 ...... Computational Science Graduate Fellow

M.S., Astronomy, The Ohio State University, 2016 ...... Columbus, OH, USA

Allan Markowitz Award in Observational Astronomy, 2017 ...... The Ohio State University, Columbus, OH, USA

viii Publications

Research Publications

1. T. W.-S. Holoien, et al., “The ASAS-SN Bright Supernova Catalog – III. 2016”, arXiv:1704.02320, (2017).

2. T. W.-S. Holoien, P. J. Marshall, and R. H. Wechsler, “EmpiriciSN: Re- Sampling Observed Supernova/Host Galaxy Populations Using an XD Gaussian Mixture Model”, AJ, 153, 249, (2017).

3. T. W.-S. Holoien, et al., “The ASAS-SN Bright Supernova Catalog – II. 2015”, MNRAS, 467, 1098, (2017).

4. T. W.-S. Holoien, et al., “The ASAS-SN Bright Supernova Catalog – I. 2013 − 2014”, MNRAS, 464, 2672, (2017).

5. T. W.-S. Holoien, et al., “ASASSN-15oi: A Rapidly Evolving, Luminous Tidal Disruption Event at 216 Mpc”, MNRAS, 463, 3813, (2016).

6. T. W.-S. Holoien, et al., “Discovery and Observations of the Unusually Luminous Type-Defying Type II-P/II-L Supernova ASASSN-13co”, Acta Astronom- ica, 66, 219, (2016).

7. T. W.-S. Holoien, et al., “Six Months of Multi-Wavelength Follow-up of the Tidal Disruption Candidate ASASSN-14li and Implied TDE Rates from ASAS-SN”, MNRAS, 455, 2918, (2016).

8. T. W.-S. Holoien, et al., “ASASSN-14ae: A Tidal Disruption Event at 200 Mpc”, MNRAS, 445, 3263, (2014).

9. T. W.-S. Holoien, et al., “Discovery and Observations of ASASSN-13db, an EX Lupi-Type Accretion Event on a Low-Mass T Tauri Star”, ApJL, 785L, 35, (2014).

10. S. Mathur, et al., “Space Telescope and Optical Reverberation Mapping Project. VII. Understanding the UV anomaly in NGC 5548 with X-Ray Spec- troscopy”, arXiv:1704.06345, (2017).

11. J. S. Brown, T. W.-S. Holoien, et al., “The Ultraviolet Spectroscopic Evolution of the Low-Luminosity Tidal Disruption Event iPTF16fnl”, arXiv:1704.02321, (2017).

ix 12. M. G. Aartsen, et al. (incl. T. W.-S. Holoien), “Multiwavelength Follow-up of a Rare IceCube Neutrino Multiplet”, arXiv:1702.06131, (2017).

13. H. P. Osborn, et al. (incl. T. W.-S. Holoien), “Periodic Eclipses of the Young Star PDS 110 Discovered with WASP and KELT Photometry”, arXiv:1705.10346, (2017).

14. M. M. Fausnaugh, et al. (incl. T. W.-S. Holoien), “Reverberation Map- ping of Optical Emission Lines in Five Active Galaxies”, ApJ, 840, 97, (2017).

15. J. E. Rodriguez, et al. (incl. T. W.-S. Holoien), “The Mysterious Dim- mings of the T Tauri Star V1334 Tau”, ApJ, 836, 209, (2017).

16. L. Pei, et al. (incl. T. W.-S. Holoien), “Space Telescope and Optical Reverberation Mapping Project. V. Optical Spectroscopic Campaign and Emission- Line Analysis for NGC 5548”, ApJ, 837, 131, (2017).

17. M. A. Gully-Santiago, et al. (incl. T. W.-S. Holoien), “Placing the Spotted T Tauri Star LKCA 4 on an HR Diagram”, ApJ, 836, 200, (2017).

18. C. S. Kochanek, et al. (incl. T. W.-S. Holoien), “Supernova Progeni- tors, Their Variability, and the Type IIP Supernova ASASSN-16fq in M66”, MNRAS, 467, 3347, (2017).

19. J. S. Brown, T. W.-S. Holoien, et al., “The Long-Term Evolution of ASASSN-14li”, MNRAS, 466, 4904, (2017).

20. D. Godoy-Rivera, et al. (incl. T. W.-S. Holoien), “The Unexpected, Long-Lasting, UV Rebrightening of the Super-Luminous Supernova ASASSN-15lh”, MNRAS, 466, 1428, (2017).

21. C. Romero-Ca˜nizales, et al. (incl. T. W.-S. Holoien), “The TDE ASASSN-14li and its Host Resolved at Parsec Scales with the EVN”, ApJL, 832L, 10, (2016).

22. C. Littleﬁeld, et al. (incl. T. W.-S. Holoien), “Return of the King: Time-Series Photometry of FO Aquarii’s Initial Recovery from its Unprecedented 2016 Low State”, ApJ, 833, 93, (2016).

23. G. J. Herczeg, et al. (incl. T. W.-S. Holoien), “The Eruption of the Candidate Young Star ASASSN-15qi”, ApJ, 831, 133, (2016).

24. J. E. Rodriguez, et al. (incl. T. W.-S. Holoien), “DM Ori: A Young

x Star Occulted by a Disturbance in its Protoplanetary Disk”, ApJ, 831, 74, (2016).

25. J. L. Prieto, et al. (incl. T. W.-S. Holoien), “MUSE Reveals a Recent Merger in the Post-starburst Host Galaxy of the TDE ASASSN-14li”, ApJL, 830L, 74, (2016).

26. J. S. Brown, B. J. Shappee, T. W.-S. Holoien, et al., “Hello Darkness My Old Friend: The Fading of the Nearby TDE ASASSN-14ae”, MNRAS, 462, 3993, (2016).

27. S. J. Schmidt, et al. (incl. T. W.-S. Holoien), “ASASSN-16ae: A Pow- erful White-Light Flare on an Early-L Dwarf”, ApJL, 828L, 22, (2016).

28. B. J. Shappee, A. L. Piro, T. W.-S. Holoien, et al., “The Young and Bright Type Ia Supernova ASASSN-14lp: Discovery, Early-time Observations, First-light Time, Distance to NGC 4666, and Progenitor Constraints”, ApJ, 826, 144, (2016).

29. M. Nicholl, et al. (incl. T. W.-S. Holoien), “SN 2015bn: A Detailed, Multi-Wavelength View of a Nearby Superluminous Supernova”, ApJ, 826, 39, (2016).

30. S. Dong, et al. (incl. T. W.-S. Holoien), “ASASSN-15lh: A Highly Super-Luminous Supernova”, Science, 351, 257, (2016).

31. A. U. Abeysekara, et al. (incl. T. W.-S. Holoien), “Gamma-Rays from the Quasar PKS 1441+25: Story of an Escape”, ApJL, 815L, 22, (2015).

32. A. Pastorello, et al. (incl. T. W.-S. Holoien), “Massive Stars Exploding in a He-rich Circumstellar Medium – VII. The Metamorphosis of ASASSN-15ed from a Narrow Line Type Ibn to a Normal Type Ib Supernova”, MNRAS, 453, 3649, (2015).

33. H. C. Campbell, et al. (incl. T. W.-S. Holoien), “Total Eclipse of the Heart: The AM CVn Gaia14aae/ASASSN-14cn”, MNRAS, 452, 1060, (2015).

34. B. J. Shappee, et al. (incl. T. W.-S. Holoien), “The Man Behind the Curtain: X-rays Drive the UV through NIR Variability in the 2013 AGN Outburst in NGC 2617”, ApJ, 788, 48, (2014).

35. B. Patel, et al. (incl. T. W.-S. Holoien), “Three Gravitationally Lensed Supernovae Behind CLASH Galaxy Clusters”, ApJ, 786, 9, (2014).

xi 36. O. Graur, et al. (incl. T. W.-S. Holoien), “Type Ia Supernova Rates to Redshift 2.4 from CLASH: the Cluster Lensing And Supernova survey with Hubble”, ApJ, 783, 28, (2014).

37. S. J. Schmidt, et al. (incl. T. W.-S. Holoien), “Characterizing a Dra- matic ∆V ∼ −9 Flare on an Ultracool Dwarf Found by the ASAS-SN Survey”, ApJL, 781L, 24, (2014).

Fields of Study

Major Field: Astronomy

xii Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments ...... v

Vita ...... viii

List of Tables ...... xvi

List of Figures ...... xix

Chapter 1: Introduction ...... 1

1.1 The All-Sky Automated Survey for Supernovae ...... 1

1.2 TidalDisruptionEvents ...... 2

1.3 Supernovae ...... 5

1.4 ScopeoftheDissertation...... 7

Chapter 2: ASASSN-14ae: A Tidal Disruption Event at 200 Mpc .. 10

2.1 ObservationsandSurveyData...... 11

2.1.1 ArchivalPhotometryandSpectroscopy ...... 11

2.1.2 NewPhotometricObservations ...... 13

2.1.3 NewSpectroscopicObservations...... 15

xiii 2.2 Analysis ...... 16

2.2.1 Light Curve Analysis ...... 16

2.2.2 SEDAnalysis ...... 17

2.2.3 Spectroscopic Analysis ...... 19

2.3 Discussion...... 22

Chapter 3: Six Months of Multi-Wavelength Follow-up of the Tidal Disruption Candidate ASASSN-14li and Implied TDE Rates from ASAS-SN ...... 42

3.1 ObservationsandSurveyData...... 44

3.1.1 ArchivalData...... 44

3.1.2 NewPhotometricObservations ...... 46

3.1.3 NewSpectroscopicObservations...... 48

3.2 Analysis ...... 49

3.2.1 SEDAnalysis ...... 49

3.2.2 Spectroscopic Analysis ...... 50

3.3 TDERates ...... 54

3.4 Discussion...... 56

Chapter 4: ASASSN-15oi: A Rapidly Evolving, Luminous Tidal Disruption Event at 216 Mpc ...... 76

4.1 ObservationsandSurveyData...... 77

4.1.1 ArchivalData...... 77

4.1.2 NewPhotometricObservations ...... 79

4.1.3 NewSpectroscopicObservations...... 81

4.2 Analysis ...... 82

4.2.1 SEDAnalysis ...... 82

xiv 4.2.2 Spectroscopic Analysis ...... 85

4.2.3 ComparisonofASAS-SNTDEs ...... 88

4.3 Discussion...... 93

Chapter 5: The Unusually Luminous Type-Defying II-P/II-L Supernova ASASSN-13co ...... 110

5.1 Observations...... 111

5.1.1 PhotometricObservations ...... 112

5.1.2 SpectroscopicObservations...... 113

5.2 LightCurveFitsandAnalysis ...... 115

5.3 Discussion...... 119

Chapter 6: The ASAS-SN Bright Supernova Catalogs ...... 133

6.1 DataSamples ...... 135

6.1.1 The ASAS-SN Supernova Sample ...... 135

6.1.2 The Non-ASAS-SN Supernova Sample ...... 136

6.1.3 TheHostGalaxySamples ...... 138

6.2 AnalysisoftheSample ...... 138

6.3 ExaminationofMissedCases ...... 143

6.4 Discussion...... 148

References ...... 162

Appendix A: Follow-up Photometry ...... 171

xv List of Tables

Table 2.1 Photometry of the Host Galaxy of ASASSN-14ae...... 40

Table 2.2 Peak Absolute Magnitudes and Estimated Decline Rates of ASASSN-14ae in Swift Filters...... 40

Table2.3 BlackbodyEvolutionofASASSN-14ae...... 41

Table 3.1 Photometry of the Host Galaxy of ASASSN-14li...... 72

Table 3.2 Spectroscopic Observations of ASASSN-14li...... 73

Table 3.3 Line Luminosities for ASASSN-14li...... 74

Table 3.4 Detection Statistics for TDE Rate Calculations...... 75

Table 4.1 Photometry of the Host Galaxy of ASASSN-15oi...... 108

Table 4.2 Estimated Photometry of the Host Galaxy of ASASSN-15oi. . . 109

Table 4.3 Spectroscopic Observations of ASASSN-15oi...... 109

Table 5.1 Photospheric Velocities of ASASSN-13co...... 131

Table 5.2 PP15 Model Parameters for ASASSN-13co...... 132

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 172

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 173

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 174

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 175

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 176

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 177

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 178

xvi Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 179

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 180

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 181

Table A.1 Follow-up Photometric Measurements of ASASSN-14ae. . . . . 182

Table A.2 Swift XRTPhotometryofASASSN-14li...... 183

Table A.2 Swift XRTPhotometryofASASSN-14li...... 184

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 185

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 186

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 187

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 188

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 189

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 190

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 191

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 192

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 193

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 194

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 195

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 196

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 197

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 198

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 199

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 200

Table A.3 Follow-up Photometric Measurements of ASASSN-14li...... 201

Table A.4 Swift XRTPhotometryofASASSN-15oi...... 201

xvii Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 202

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 203

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 204

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 205

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 206

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 207

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 208

Table A.5 Follow-up Photometric Measurements of ASASSN-15oi. . . . . 209

Table A.6 Follow-up Photometric Measurements of ASASSN-13co. . . . . 210

Table A.6 Follow-up Photometric Measurements of ASASSN-13co. . . . . 211

Table A.6 Follow-up Photometric Measurements of ASASSN-13co. . . . . 212

Table A.6 Follow-up Photometric Measurements of ASASSN-13co. . . . . 213

xviii List of Figures

Figure2.1 DiscoveryimageofASASSN-14ae...... 25

Figure2.2 LightcurvesofASASSN-14ae...... 26

Figure 2.3 Observed spectral energy distribution of ASASSN-14ae..... 27

Figure 2.4 Spectral time-sequence of ASASSN-14ae...... 28

Figure 2.5 Host-galaxy-subtracted spectral time-sequence of ASASSN-14ae. 29

Figure 2.6 Swift colorevolutionofASASSN-14ae...... 30

Figure 2.7 Color comparison between ASASSN-14ae and SNe...... 31

Figure2.8 SEDﬁtsofASASSN-14ae...... 32

Figure 2.9 Blackbody temperature evolution of ASASSN-14ae...... 33

Figure 2.10 Blackbody luminosity evolution of ASASSN-14ae...... 34

Figure 2.11 Implied luminosity in Eddington units of ASASSN-14ae using athindiskmodel...... 35

Figure 2.12 Early-phase spectroscopic comparison of ASASSN-14ae to SNe andAGN...... 36

Figure 2.13 Late-phase spectroscopic comparison of ASASSN-14ae to SNe andAGN...... 37

Figure 2.14 Evolution of the Hα line proﬁle of ASASSN-14ae...... 38

Figure 2.15 SED of a thin disk ﬁt to the emission of ASASSN-14ae. . . . . 39

Figure 3.1 Discovery image of ASASSN-14li...... 61

Figure3.2 LightcurvesofASASSN-14li...... 62

Figure 3.3 Spectroscopic time-sequence of ASASSN-14li...... 63

xix Figure 3.4 Host-galaxy-subtracted spectroscopic time-sequence of ASASSN-14li...... 64

Figure 3.5 SED evolution of ASASSN-14li...... 65

Figure 3.6 Luminosity evolution of ASASSN-14li...... 66

Figure 3.7 Spectroscopic comparison of ASASSN-14li and ASASSN-14ae. 67

Figure 3.8 Evolution of ASASSN-14li’s line luminosities...... 68

Figure 3.9 Comparison of the evolution of the Hα line proﬁles of ASASSN- 14liandASASSN-14ae...... 69

Figure 3.10 Comparison of the evolution of the Hα line proﬁles of ASASSN- 14liandASASSN-14ae...... 70

Figure 3.11 Evolution of ASASSN-14li’s line luminosities...... 71

Figure4.1 DiscoveryimageofASASSN-15oi...... 95

Figure4.2 LightcurvesofASASSN-15oi...... 96

Figure 4.3 Spectroscopic time-sequence of ASASSN-15oi...... 97

Figure 4.4 Unfolded X-ray spectrum of ASASSN-15oi...... 98

Figure4.5 EvolutionoftheSEDofASASSN-15oi...... 99

Figure 4.6 Luminosity Evolution of ASASSN-15oi...... 100

Figure 4.7 Spectroscopic comparison between ASASSN-15oi and TDEs andSNe...... 101

Figure 4.8 Comparison of the (UV-UV) and (UV-optical) color evolution of ASASSN-15oi, ASASSN-14li, and ASASSN-14ae...... 102

Figure 4.9 Comparison of the blackbody SEDs of ASASSN-15oi, ASASSN- 14li,andASASSN-14ae...... 103

Figure 4.10 Comparison of the luminosity evolution of ASASSN-15oi, ASASSN-14li,andASASSN-14ae...... 104

Figure 4.11 Comparison of the temperature evolution of ASASSN-15oi, ASASSN-14li,andASASSN-14ae...... 105

xx Figure 4.12 Comparison of the photospheric radius evolution of ASASSN- 15oi,ASASSN-14li,andASASSN-14ae...... 106

Figure 4.13 Comparison of the evolution of the FWHM of prominent emission lines in ASASSN-15oi, ASASSN-14li, and ASASSN-14ae. . . 107

Figure5.1 DiscoveryimageofASASSN-13co...... 121

Figure5.2 LightcurvesofASASSN-13co...... 122

Figure 5.3 Spectroscopic time-sequence of ASASSN-13co...... 123

Figure 5.4 Spectroscopic comparison between ASASSN-13co and Type II- PSNe...... 124

Figure 5.5 Velocity evolution of the emission lines of ASASSN-13co. . . . 125

Figure 5.6 PP15 Model light curve ﬁts of ASASSN-13co...... 126

Figure 5.7 Evolution of the bolometric luminosity of ASASSN-13co comparedtoPP15SNe...... 127

Figure 5.8 Comparison of the absolute V -band light curve of ASASSN-13co totheAndersonetal.(2014)SNIIsample...... 128

Figure 5.9 Comparison of the absolute I-band light curve of ASASSN-13co totheOGLESNIIsample...... 129

Figure 5.10 Correlation of the absolute V -band magnitude and the V -band light curve slope measured at τ = 0 for ASASSN-13co and the PP15 SNIIsample...... 130

Figure 6.1 Breakdown by type of the SNe discovered by ASAS-SN between 2014May01and2016December31...... 150

Figure 6.2 Breakdown by type of the SNe discovered by non-ASAS-SN sources between 2014 May 01 and 2016 December 31...... 151

Figure 6.3 Breakdown by type of the full combined bright SN sample. . . 152

Figure 6.4 Oﬀset from host nucleus in arcseconds vs. absolute KS-band magnitude for all SNe in the combined SN sample...... 153

Figure 6.5 Oﬀset from host nucleus in kpc vs. absolute KS-band magnitude for all SNe in the combined SN sample...... 154

xxi Figure 6.6 Cumulative, normalized distributions of host galaxy absolute magnitudes for all SNe in the combined SN sample...... 155

Figure 6.7 Oﬀset from host nucleus in kpc vs. absolute KS-band magnitude for all SNe in the combined SN sample...... 156

Figure 6.8 Histogram of bright SN discoveries in each month from 2012 through2016...... 157

Figure 6.9 Cumulative normalized histogram of SN discoveries with respect tothesineoftheirdeclination...... 158

Figure 6.10 Histograms of SN redshifts from our complete sample...... 159

Figure 6.11 Cumulative histograms of SN peak magnitudes from our completesample...... 160

Figure 6.12 Image of OGLE-2014-SN-067, which was not recovered in ASAS-SNdata...... 161

xxii Chapter 1: Introduction

1.1. The All-Sky Automated Survey for Supernovae

While the night sky may look constant to the naked eye, it is in fact filled with transient and variable events that help shape our universe. These energetic, violent events can help us measure the universe and probe physical conditions too extreme to test on Earth. Because of this, there have been numerous large-scale sky surveys designed to scan large areas of the sky repeatedly to find objects that change over time, and upcoming surveys like the Large Synoptic Survey Telescope (LSST; Ivezic et al. 2008) represent one of the major branches of future astronomy outlined in the most recent Astronomy and Astrophysics Decadal Survey (National Research Council 2010). However, despite the number of such survey projects, there was no optical survey designed to scan the entire visible night sky on a rapid cadence to find the bright, nearby transients that can be studied in the greatest detail and have the greatest impact on our understanding of these violent events. This changed in 2013 with the creation of the All-Sky Automated Survey for SuperNovae (ASAS-SN; Shappee et al. 2014).

ASAS-SN is a long-term project designed to monitor the entire sky on a rapid cadence to find nearby supernovae and other bright transients, such as tidal disruption events, flares from active galactic nuclei (AGN), and stellar outbursts. This is accomplished using telescopes with 14-cm aperture lenses and standard V -band filters, giving a 4.5 × 4.5 degree field-of-view and a limiting magnitude of mV ∼ 17. Data are downloaded, reduced, and searched in real-time, allowing for rapid discovery and response.

ASAS-SN began its real-time sky survey in 2013 April with our ﬁrst unit, Brutus, consisting of two telescopes on a common mount hosted at the Las Cumbres

1 Observatory Global Telescope Network (LCO; Brown et al. 2013) site on Mount Haleakala, Hawaii. In late 2013, Brutus was upgraded with two additional cameras, giving a sky coverage of roughly 10,000 square degrees per clear night. In the spring of 2014 we deployed our second unit, Cassius, again consisting of two telescopes on a common mount, at the LCO site at Cerro Tololo, Chile. Cassius began on-sky operations on 2014 May 1, and we consider this the oﬃcial start date of the two-hemisphere ASAS-SN. Cassius was upgraded to four telescopes in 2015 July, and ASAS-SN now covers roughly 16,000 square degrees per clear night, and covers the entire observable sky (∼ 30, 000 square degrees on a given night) with a cadence of 2 − 3 days.

ASAS-SN data are processed and searched in real-time and all ASAS-SN discoveries are announced publicly upon confirmation, allowing for rapid discovery and response by both the ASAS-SN team and others. Further, ASAS-SN uses an untargeted survey approach: rather than focusing on a specific list of galaxies or small region of the sky, ASAS-SN surveys the entire sky, prioritizing observations to minimize the amount of time between visits of each observed field. This allows ASAS-SN to produce less-biased samples of a wide variety of bright transient and variable events that are ideal for population studies. Since all ASAS-SN discoveries are bright, they can be easily observed with modest follow-up resources, and this has allowed us to produce detailed datasets for many of the most interesting individual ASAS-SN discoveries.

In this dissertation, I use ASAS-SN discoveries to investigate the observational and theoretical properties of two types of transient events caused by the explosive deaths of stars: tidal disruption events (TDEs) and supernovae (SNe).

1.2. Tidal Disruption Events

When a star’s orbit brings it within the tidal disruption radius of a supermassive black hole (SMBH), the tidal shear forces become more powerful than the self-gravity of the star and the star is torn apart. Roughly half of the mass of the star is ejected while the rest of the stellar material remains bound to the black hole. As the bound material returns to the black hole various dynamical processes likely cause most of

2 the material to eventually become unbound, with a small fraction accreting onto the SMBH. These “TDEs” are expected to produce short-lived (t ∼< 1 yr), luminous accretion flares, releasing more energy than even SNe (Lacy, Townes & Hollenbach 1982; Phinney 1989; Rees 1988; Evans & Kochanek 1989). For black hole masses 7 MBH ∼< 10 M⊙, the initial fallback rate is super-Eddington and the eventual rate at which material returns to pericenter becomes a t−5/3 power law (Evans & Kochanek 1989; Phinney 1989). It was assumed that the observed luminosity of the TDE flare would be proportional to this rate of return of the stellar material to pericenter, and that the luminosity of the flare would peak at soft X-ray energies.

Since they were theoretically predicted in the late 1980s, roughly 20 likely TDE candidates have been observed. These observed TDE candidates have exhibited a large diversity in properties. TDE flares have now been discovered at hard X-ray (e.g., Bloom et al. 2011; Burrows et al. 2011; Cenko et al. 2012a; Pasham et al. 2015), soft X-ray (e.g., Bade, Komossa & Dahlem 1996; Grupe, Thomas & Leighly 1999; Komossa & Greiner 1999; Donley et al. 2002; Maksym, Ulmer & Eracleous 2010; Saxton et al. 2012), ultraviolet (UV; e.g., Stern et al. 2004; Gezari et al. 2006, 2008, 2009), and optical (e.g., van Velzen et al. 2011; Gezari et al. 2012; Cenko et al. 2012b; Arcavi et al. 2014; Chornock et al. 2014; Holoien et al. 2014; Vinkóet al. 2015; Holoien et al. 2016a; Holoien et al.,Holoien et al. 2016c) wavelengths—see Komossa (2015) for a review. Some TDE candidates appear to follow the predicted t−5/3 power law decline, but others exhibit different decline rates (e.g., Vinkóet al. 2015; Holoien et al. 2016a; Brown et al. 2016). Many of the X-ray bright candidates do not show associated strong optical emission, and many of the recent candidates discovered by optical surveys show strong optical and UV emission without associated X-ray emission, with the notable exception of ASASSN-14li, which was discovered in the optical (Holoien et al. 2016a) but also detected at both X-ray and radio wavelengths (Miller et al. 2015; Alexander et al. 2016; van Velzen et al. 2016; Holoien et al. 2016a). The optically discovered TDEs also exhibit a continuum of spectroscopic properties ranging from He-dominated to H-dominated (Arcavi et al. 2014). This diversity could be due to photoionization physics (e.g., Guillochon, Manukian & Ramirez-Ruiz 2014; Gaskell & Rojas Lobos 2014; Roth et al. 2016), composition variations created by stellar initial composition and evolution (Kochanek 2016a),

3 and/or require the disruption of some helium stars (Gezari et al. 2012; Strubbe & Murray 2015).

The properties of TDE flares depend on numerous physical parameters, including the depth of the encounter, the composition of the star, the fraction of the star that is accreted, and the geometry of the accretion stream (e.g., Kochanek 1994; Lodato & Rossi 2011; Guillochon & Ramirez-Ruiz 2015; Metzger & Stone 2016; Shiokawa et al. 2015). If the TDE flare is powered by the accretion of the disrupted stellar material onto the black hole, the observed optical/UV emission is likely dominated by a photosphere formed within the stellar debris (Evans & Kochanek 1989; Loeb & Ulmer 1997; Ulmer 1999; Strubbe & Quataert 2009; Roth et al. 2016), and is likely dependent on the viewing angle (Guillochon, Manukian & Ramirez-Ruiz 2014; Metzger & Stone 2016). If the flare is powered by interactions within the debris stream during disk formation, the observed emission depends strongly on the dynamics of the system, the outflow of unbound material, and on the orientation of the debris with respect to the observer (Piran et al. 2015; Shiokawa et al. 2015; Svirski, Piran & Krolik 2017).

A better understanding of the physics behind TDEs may improve our understanding of galaxy and SMBH evolution—despite their relatively low frequency (10−5 − 10−4 yr−1 per galaxy; van Velzen et al. 2011; Holoien et al. 2016a), TDEs may have an influence on the evolution of SMBH spin and mass in their host galaxies. Further, the optically discovered TDEs have shown a preference for E+A galaxies, which show little-to-no line emission in their spectra but exhibit strong Balmer line absorption, indicating that the galaxy recently went through a period of intense star formation (French, Arcavi, & Zabludoff 2016). TDEs may play an even more important role in the evolution of these galaxies. TDEs allow us to study SMBHs in otherwise quiescent galaxies, as the emission from the TDE flare is sensitive to black hole mass and spin (e.g., Ulmer 1999; Graham et al. 2001). Only through the discovery and detailed observation of additional TDEs can we being to fully understand the processes governing the observed emission of TDE flares, and the nearby TDEs discovered by ASAS-SN have proven to be ideal candidates for detailed study.

4 1.3. Supernovae

SNe—the explosive deaths of stars—are some of the most energetic transient events in the universe, capable of outshining the combined light of billions of other stars in their host galaxies for weeks. However, they are diﬃcult to study due to their rarity and unpredictability, with roughly one occurring every 100 years in galaxies similar to our own. The Milky Way has not hosted a SN in over 400 years, but through dozens of survey projects and SN searches like ASAS-SN, astronomers have now discovered tens of thousands of SNe. Most of these were discovered by projects designed to use SNe for cosmology, and thus the majority of known SNe were found in distant galaxies, preventing the kind of in-depth study needed to investigate the physics behind these violent events. Because of this, SNe remain poorly understood in many ways, and the detailed study of bright, nearby SNe present the best method of furthering our understanding.

SNe are divided into two main classes: thermonuclear and core-collapse. Type Ia SNe (SNe Ia) are thought to be thermonuclear explosions of carbon-oxygen white dwarfs (WDs) and are the most luminous SNe outside of the rare “superluminous” subclass. The peak luminosities and luminosity decline rates of SNe Ia are well-correlated, and this allows them to be used as “standardizeable candles” to measure cosmological distances (e.g., Phillips 1993; Hamuy et al. 1995; Riess et al. 1995). Distances measured using SNe Ia were used in the discovery that the universe is accelerating, which was awarded the 2011 Nobel Prize in Physics (Riess et al. 1998; Perlmutter et al. 1999). However, as many thousands of SNe Ia have been discovered, systematic errors in our understanding of these events now dominate the uncertainty in cosmological measurements made using SNe Ia (e.g., Guy et al. 2010; Conley et al. 2011).

A major gap in our understanding of SNe Ia is the nature of their progenitor systems (see Wang & Han 2012 for a review). SNe Ia are thought to be the termonuclear explosion of WDs in binary systems, but it is unknown whether the binary companion is a non-degenerate object, like a main sequence star or red giant (the “single degenerate” model, Whelan & Iben 1973; Nomoto 1982) or whether it is another WD (the “double degenerate” model, Tutukov & Yungelson 1979; Iben &

5 Tutukov 1984; Webbink 1984). There are a few ways in which the progenitor model of SNe Ia can be investigated. One is through the detailed study of individual SNe Ia. In particular, the very early (t ∼< 1 day) and late-time (t ∼> a few years) luminosity evolution is sensitive to the density and composition of the binary companion (e.g., Shappee et al. 2016a,b, 2017). The second is through the study of the rate at which SNe Ia occur. The distribution of times between episodes of star formation and subsequent SNe Ia is sensitive to the SN Ia progenitor model, and we can calculate this distribution using SN Ia rates and star formation histories (e.g., Wang & Han 2012; Graur et al. 2014). Study of early- and late-time light curves requires very bright SNe, and the calculation of accurate SN rates requires a sample of SNe that is as unbiased as possible. Thus, as survey that uses rapid-cadence, untargeted observations of the entire sky to search for bright transients, such as ASAS-SN, is an ideal tool for studying the progenitors of SNe Ia.

Core-collapse SNe (CCSNe) are more common and typically less luminous than SNe Ia, and are further subdivided into those with hydrogen in their spectra (Type II) and those without (Type I). Type II SNe (SNe II) have been widely studied and are known to arise from progenitors that retain their hydrogen envelopes before exploding as CCSNe (see Filippenko 1997 for a review). SNe II have traditionally been separated into two groups based on the shapes of their light curves: Type II-Linear (II-L) SNe show a steady decline in magnitude at optical wavelengths, while Type II-Plateau (II-P) SNe exhibit a lengthy “plateau” phase in their optical light curves during which the magnitude decline is very small (e.g., Arcavi et al. 2012; Faran et al. 2014a,b). These observed differences are thought to arise from differences in the thickness of the hydrogen envelope remaining around the star (e.g., Sanders et al. 2015). However, it recent years it has been suggested that this distinction is perhaps an oversimplification, and that SN II light curve shapes are simply a continuum of properties (e.g., Anderson et al. 2014; Sanders et al. 2015).

Determining whether there are indeed two distinct classes of SNe II or whether they fall into a broader continuum will tell us about the properties of their progenitor stars and the physics driving the emission we observe from the SNe. Investigating the light curves of SNe II can be done through the detailed study of individual SNe that

6 do not fall into the typical II-P or II-L classiﬁcation system or through the statistical analysis of a large sample of well-observed SNe, ideally one that is obtained with minimal observational bias and with consistent ﬂux measurements. ASAS-SN again represents an ideal tool for these types of studies, as all of its discoveries are bright and nearby, and its sample continues to grow every year the survey continues.

There are other, rarer SNe that I do not discuss in detail here, including Type Ib/c, Type IIn, Type IIb, and superluminous SNe (SLSNe). Though these are fewer in number, ASAS-SN has discovered examples of each of these classes of SNe. ASAS-SN was designed primarily as a survey for SNe, and with hundreds of bright SNe discovered to date it continues to be an ideal tool for examining SNe in ways that would not be possible with a fainter, more distant sample.

1.4. Scope of the Dissertation

In Chapter 2 I present ground-based and Swift follow-up photometric and spectroscopic observations of ASASSN-14ae, the ﬁrst TDE discovered by ASAS-SN, at a distance of roughly 200 Mpc. We ﬁnd that the transient had a peak luminosity of L ≃ 8×1043 erg s−1 and that it radiated a total integrated energy of E ≃ 1.7×1050 ergs over the ∼ 5 months of observations presented. The blackbody temperature of the transient remains roughly constant at T ∼ 20, 000 K while the luminosity declines by nearly 1.5 orders of magnitude during this time, a drop that is most −t/t0 consistent with an exponential decline, L ∝ e with t0 ≃ 39 days, rather than the power-law decline predicted by theory. We compare the color and spectroscopic evolution to both SNe and normal AGN to show that ASASSN-14ae does not resemble either type of object and conclude that a TDE is the most likely explanation for our observations. At z =0.0436, ASASSN-14ae was the nearest TDE candidate ever discovered at the time of discovery.

In Chapter 3 I present ∼ 6 months of photometric and spectroscopic observations of ASASSN-14li, the brightest TDE ever discovered and the second TDE found by ASAS-SN. ASASSN-14li had a peak bolometric luminosity of L ≃ 1044 ergs s−1 and a total integrated energy of E ≃ 7 × 1050 ergs radiated over the time period shown. The UV/optical emission of the source is well-ﬁt by

7 a blackbody with roughly a constant temperature of T ∼ 35, 000 K, while the luminosity declines by roughly a factor of 16 over this time. As was the case with ASASSN-14ae, the optical/UV luminosity decline is broadly consistent with an −t/t0 exponential decline, L ∝ e , with t0 ≃ 60 days. ASASSN-14li also exhibited soft X-ray emission comparable in luminosity to the optical and UV emission but declining at a slower rate, and the X-ray emission dominated in later epochs. We use the discoveries of ASASSN-14li and ASASSN-14ae to estimate the TDE rate implied by ASAS-SN, finding an average rate of r ≃ 5.4 × 10−5 yr−1 per galaxy with a 90% confidence interval of (2.2 − 17.0) × 10−5 yr−1 per galaxy, which is consistent within uncertainties with both previously established theoretical and observational rates. As the brightest and one of the longest-lasting TDE flares ever discovered, ASASSN-14li now is the most well-studied TDE, with hundreds of observations spanning hard X-ray through radio wavelengths.

In Chapter 4 I present ground-based and Swift observations of the ASASSN- 15oi, the third TDE discovered by ASAS-SN, which was unusually luminous and evolved at an atypically rapid rate. ASASSN-15oi peaked at a bolometric luminosity of L ≃ 1.3 × 1044 ergs s−1 and radiated a total energy of E ≃ 6.6 × 1050 ergs over the ∼ 3.5 months of observations shown. We fit the early optical/UV emission of the source with a blackbody with temperature increasing from T ∼ 2 × 104 K to T ∼ 4 × 104 K over the course of the observations. The luminosity declines from L ≃ 1.3 × 1044 ergs s−1 to L ≃ 2.3 × 1043 ergs s−1, requiring the photosphere to be shrinking rapidly. We again find that the optical/UV luminosity decline during this period is most consistent with an exponential decline, L ∝ e−(t−t0)/τ , with −α τ ≃ 46.5 days for t0 ≃ 57241.6 (MJD), while a power-law decline of L ∝ (t − t0) with t0 ≃ 57212.3 and α = 1.62 provides a moderately worse fit. ASASSN-15oi also exhibits roughly constant soft X-ray emission that is significantly weaker than the optical/UV emission. The early spectroscopic features and color evolution of ASASSN-15oi are consistent with a TDE, but the rapid spectral evolution is unique among optically-selected TDEs.

In Chapter 5 I present observations of ASASSN-13co, an unusually luminous SN II and the ﬁrst CCSN discovered by ASAS-SN. The SN was ﬁrst detected on

8 2013 August 29 and the data presented span roughly 3.5 months after discovery. We use the recently developed model from (Pejcha & Prieto 2015a) to model the multi-band light curves of ASASSN-13co and derive the bolometric luminosity curve. We compare ASASSN-13co to other SNe II to show that it was unusually luminous and that it exhibited an atypical light curve shape that neither cleanly matches that of a standard Type II-L nor Type II-P SN, providing further evidence that SNe II span a range of light curve properties rather than falling into two distinct classes.

Finally, in Chapter 6 I describe the full sample of all SNe discovered by ASAS-SN during its ﬁrst three-and-a-half years of operations, spanning 2013,

2014, 2015, and 2016. I also discuss the sample of all other bright (mV ≤ 17), spectroscopically conﬁrmed SNe discovered from 2014 May 1 through the end of 2016, providing a comparison to the ASAS-SN sample starting from the point where ASAS-SN became operational in both hemispheres. The cumulative combined sample now totals 668 bright SNe discovered since 2014 May 1, and I provide statistical analyses of the SNe and their host galaxies from our combined sample. This bright SN sample allows for population studies that were not previously possible because the all-sky emphasis of ASAS-SN redresses most previously existing biases. In particular, ASAS-SN systematically ﬁnds supernovae closer to the centers of host galaxies than either other professional surveys or amateurs, a remarkable result given ASAS-SN’s poorer angular resolution.

9 Chapter 2: ASASSN-14ae: A Tidal Disruption Event at 200 Mpc

In this chapter, I describe the discovery and follow-up observations of ASASSN-14ae, the ﬁrst TDE discovered by ASAS-SN, and at the time the nearest TDE ever discovered.

The ASAS-SN transient source detection pipeline was triggered on 2014 January 25, detecting a new source with V = 17.1 ± 0.1 mag (Prieto et al. 2014). The object was also detected on 2014 January 26 at roughly the same magnitude, but is not detected (V ∼> 18 mag) in data obtained on 2014 January 1 and earlier. A search at the object’s position in the Sloan Digital Sky Survey Data Release 9 (SDSS DR9; Ahn et al. 2012) catalog revealed the source of the outburst to be the inclined spiral galaxy SDSS J110840.11+340552.2 at redshift z =0.0436, corresponding to a −1 −1 luminosity distance of d = 193 Mpc (H0 = 70 km s Mpc , ΩM =0.3, ΩΛ =0.7), and that the ASAS-SN source position was consistent with the center of the host galaxy. Follow-up images obtained on 2014 January 27 with the LCO 1-m telescope at McDonald Observatory (Brown et al. 2013), the 2-m Liverpool Telescope (LT) (Steele et al. 2004), and the Swift UltraViolet and Optical Telescope (UVOT; Roming et al. 2005) conﬁrmed the detection of the transient. After astrometrically aligning an LT image of the source in outburst with the archival SDSS image of the host galaxy, we measured an oﬀset of 0.28 ± 0.45 pixels (0.09 ± 0.14 arcseconds) between the position of the brightest pixel in the host galaxy in the LT image and the position of the brightest pixel in the SDSS image. This indicated that the source

This chapter is adapted from “ASASSN-14ae: A Tidal Disruption Event at 200 Mpc”, T. W.-S. Holoien, et al. MNRAS, 445, 3263, (2014).

10 of the new ﬂux was consistent with the center of the galaxy. Figure 2.1 shows the ASAS-SN V -band reference and subtracted images of the source as well as SDSS pre-discovery and LT g-band images.

The archival SDSS spectrum of the host is that of an early-type spiral with little evidence of emission lines from an AGN, although it does show [O III] 5007 in emission indicating that there is some recent star-formation. A transient classiﬁcation spectrum obtained on 2014 January 29 with the Dual-Imaging Spectrograph (DIS) mounted on the Apache Point Observatory (APO) 3.5-m telescope showed a blue continuum as well as a broad (FWHM ≃ 17000 km/s) Hα line. The blue continuum and Hα emission suggested that this transient was likely a young SN II, but the proximity to the galactic nucleus and its absolute magnitude at discovery (MV ∼ −19.3 mag from the ASAS-SN host-subtracted image) made a TDE a potential alternative. We decided to start a follow-up campaign in order to fully characterize this interesting transient.

In §2.1 I describe pre-outburst archival observations, including both photometry and spectroscopy of the host galaxy, as well as new data taken of the transient during our follow-up campaign. In §2.2 I analyze these data and describe the properties of the transient. Finally, in §2.3 I compare these properties to those of SNe, AGN, and other candidate TDEs to examine the nature of ASASSN-14ae and discuss late-time observations of the object that were obtained years after discovery.

2.1. Observations and Survey Data

In this section I summarize the available archival data of the transient host galaxy as well as our new photometric and spectroscopic observations of ASASSN-14ae.

2.1.1. Archival Photometry and Spectroscopy

We retrieved archival reduced images in ugriz of SDSS J110840.11+340552.2 from SDSS DR9. We then measured the ﬂuxes in a 5′′.0 aperture radius (the same aperture used to measure the source in follow-up data, chosen to match the Swift point spread function (PSF) and to minimize the eﬀects of seeing variations on the photometry) to use for galaxy spectral energy distribution (SED) modeling and for subtracting

11 the host galaxy ﬂuxes from the transient ﬂuxes. We also retrieved near-IR JHKs images from the Two-Micron All Sky Survey (2MASS; Skrutskie et al. 2006) and measured aperture magnitudes of the host galaxy in the same fashion. The measured SDSS and 2MASS magnitudes of the host galaxy are presented in Table 2.1.

There are no archival Spitzer, Herschel, Hubble Space Telescope (HST), Chandra, or X-ray Multi-Mirror Mission (XMM-Newton) observations of the source. The host galaxy is not detected in the ROSAT All-Sky Survey (RASS; Voges et al. 1999) with an upper ﬂux limit of 3 × 10−13 erg s−1 cm−2 in the 0.1 − 2.4 keV band, providing further evidence that the galaxy is inactive. We also retrieved archival mid-IR photometry from the Wide-ﬁeld Infrared Survey Explorer (WISE; Wright et al. 2010). From the WISE W 1 and W 2 measurements we calculate that the host galaxy has (W 1 − W 2) ≃ 0.06 ± 0.06 mag, and this blue mid-IR color is further evidence against AGN activity (e.g., Assef et al. 2013).

We used the code for Fitting and Assessment of Synthetic Templates (Fast v1.0; Kriek et al. 2009) to ﬁt stellar population synthesis (SPS) models to the ′′ 5.0 SDSS ugriz and 2MASS JHKs magnitudes of the host galaxy. The ﬁt was made assuming a CCM extinction law (Cardelli, Clayton & Mathis 1988) with

RV = 3.1, an exponentially declining star-formation history, a Salpeter initial mass function, and the Bruzual & Charlot (2003) models. We obtained a good 2 +0.15 SPS fit (reduced χν = 0.4), with the following parameters: AV = 0.15−0.15 mag, +0.6 9 +0.6 +5 −2 −1 M∗ = (6.3−0.8) × 10 M⊙, age= 2.2−0.2 Gyr, and SFR= (3−1) × 10 M⊙ yr . These properties do not appear to be consistent with SDSS J110840.11+340552.2 being an E+A galaxy, as many of the hosts of the TDE candidates in Arcavi et al. (2014) were. The Fast estimate of AV = 0.15 mag incorporates both Galactic and host extinction and this value is consistent with the Galactic extinction (AV =0.057 mag based on Schlafly & Finkbeiner 2011). In fits to the transient SED, we find no evidence for additional extinction, even though the Swift UV data, particularly the UV M2 band which lies on top of the 2200 A˚ extinction curve feature, is a powerful probe for additional dust (though this depends on the strength of the UV bump in the dust law). In the analyses of the event’s SED which follow we only correct for this Galactic extinction.

12 We also obtained the spectrum of SDSS J110840.11+340552.2 from SDSS-DR9. The archival spectrum is dominated by absorption lines (e.g., Balmer lines, Ca I G-band, Mg I, Na I, Ca H&K, and the 4000 A˚ break) that are characteristic of intermediate-age and old stellar populations. This is consistent with the results of the Fast ﬁt to the SED of the host. The spectrum does not show strong emission lines, except for the detection of an unresolved [O III] 5007 line with FWHM ≃ 250 km s−1 39 −1 and integrated luminosity L[OIII] ≃ 2.4 × 10 erg s . This is likely a sign of a low level of recent star formation, indicating the galaxy could host core-collapse SN events, but without detecting other emission lines (e.g., Hβ, Hα, [N II]) we cannot constrain the rate of star formation. We note that the [O III]/Hβ and [N II]/Hα ratios may indicate that the host contains a very weak Type 2 AGN, consistent with the analysis done in Arcavi et al. (2014). However, other factors (e.g., the WISE photometry) argue against signiﬁcant activity.

We use the FWHM ≃ 250 km s−1 of the [O III] line from the latest spectroscopic epoch, presented in §2.1.3, to estimate a velocity dispersion for the galaxy and 6.8 place an upper limit on the mass of the black hole of MBH < 10 M⊙ using the M-σ relation from Gültekin et al. (2009). SDSS reports a velocity dispersion of σ = 42.9 ± 7.3 km s−1, which is well below the velocity resolution of the SDSS spectrograph of ∼ 100 km s−1, so we will conservatively regard the SDSS estimate as a limit of σ < 100 km s−1. Using the same M-σ relation from Gültekin et al. (2009) −1 6.9 and the SDSS resolution of 100 km s gives an upper limit of MBH ∼< 10 M⊙, consistent with the limit we derive from the width of the [O III] line. Finally, from 9.8 the Fast fit, we have M∗ ∼ 10 M⊙, which is consistent with a bulge mass of 9.4 MB ∼ 10 M⊙ (Mendel et al. 2014). Using the MB-MBH relation from McConnell 6.8 & Ma (2013) gives MBH ∼ 10 M⊙, which is again consistent with the limits derived from the host and transient spectra.

2.1.2. New Photometric Observations

After detection of the transient, we were granted a series of Swift X-ray Telescope (XRT; Burrows et al. 2005) and UVOT target-of-opportunity (TOO) observations. The Swift UVOT observations of ASASSN-14ae were obtained in 6 ﬁlters: V

13 (5468 A),˚ B (4392 A),˚ U (3465 A),˚ UV W 1 (2600 A),˚ UV M2 (2246 A),˚ and UV W 2 (1928 A)˚ (Poole et al. 2008). We used the UVOT software task Uvotsource to extract the source counts from a 5′′.0 radius region and a sky region with a radius of ∼40′′.0. The UVOT count rates were converted into magnitudes and ﬂuxes based on the most recent UVOT calibration (Poole et al. 2008; Breeveld et al. 2010). The UVOT Vega magnitudes are shown along with other photometric data in Figure 2.2.

The XRT was operating in Photon Counting mode (Hill et al. 2004) during our observations. The data from all epochs were reduced and combined with the software tasks Xrtpipeline and Xselect to obtain an image in the 0.3−10 keV range with a total exposure time of ∼ 42, 030 s. We used a region with a radius of 20 pixels (47′′.1) centered on the source position to extract source counts and a source-free region with a radius of 100 pixels (235′′.7) for background counts. We do not detect X-ray emission from ASASSN-14ae to a 3-sigma upper limit of 5.9 × 10−4 counts s−1. To convert this to a ﬂux, we assume a power law spectrum with Γ = 2 and Galactic H I column density (Kalberla et al. 2005), yielding an upper limit of ∼ 2.9 × 10−14 erg cm−2 s−1. At the host distance of d = 193 Mpc, 41 −1 7 this corresponds to an upper limit of LX ≤ 1.3 × 10 erg s (3.4 × 10 L⊙) on the average X-ray luminosity. The constraints for the individual Swift epochs are on average ∼ 10 times weaker, and we only consider the combined X-ray limit.

In addition to the Swift observations, we obtained gri images with the LCO 1-m at the MacDonald Observatory and ugriz images with the LT 2-m telescope. We measured aperture photometry1 using a 5′′.0 aperture radius to match the host galaxy and Swift UVOT measurements. The photometric zero points were determined using several SDSS stars in the ﬁeld. These data are shown in Figure 2.2.

Figure 2.2 shows the UV and optical light curves of ASASSN-14ae from MJD 56682.5 (the epoch of first detection) through our latest epoch of observations on MJD 56828 (146 days after first detection) without extinction correction or host flux subtraction. Also shown are the SDSS ugriz magnitudes and synthesized Swift

1We also attempted to do image subtraction with the SDSS archival images as templates. However, due to the lack of stars in the ﬁeld-of-view close to ASASSN-14ae, the quality of the subtractions was sub-optimal.

14 UVOT magnitudes of the host galaxy extrapolated from the host SED fit. With the host flux included, the light curve shows ASASSN-14ae brightened much more strongly in the blue and UV filters than in the red bands, with the largest increase in the Swift UV M2 band (2246 A),˚ where it brightened by ∆mUVM2 ∼ −4.9. The brightness also appears to be declining at a faster rate with respect to the host in bluer filters. We further analyze the light curve and compare it to SNe and TDEs in the literature in §2.2.1.

After correcting our photometric measurements for Galactic extinction, we construct SEDs for 5 follow-up epochs. These are shown with the extinction-corrected SDSS archival data and host SED ﬁt from Fast in Figure 2.3.

2.1.3. New Spectroscopic Observations

We obtained seven low- and medium-resolution optical spectra of ASASSN-14ae spanning more than four months between 2014 January 29 and 2014 June 6. The spectra were obtained with DIS mounted on the Apache Point Observatory 3.5-m telescope (range 3500 − 9800 A,˚ R ∼ 1000) and with the Multi-Object Double Spectrographs (MODS; Pogge et al. 2010) on the 8.4-m Large Binocular Telescope (LBT) on Mount Graham (range 3200 − 10000 A,˚ R ∼ 2000). The spectra from DIS were reduced using standard techniques in Iraf and the spectra from MODS were reduced using a custom pipeline written in Idl2. We applied telluric corrections to all the spectra using the spectrum of the spectrophotometric standard observed the same night. We calculated synthetic r-band magnitudes and scaled the ﬂuxes in each spectrum to match the r-band photometry. Figure 2.4 shows a montage of the ﬂux-calibrated spectra from both DIS and MODS, while Figure 2.5 shows the same six spectra with the host galaxy spectrum subtracted.

The main characteristics of the spectra of ASASSN-14ae are the blue continuum, consistent with the photometric measurements, and the detection of broad Balmer lines in emission, which are not present in the host galaxy spectrum. The Hα line −1 has FHWM ∼> 8000 km s at all epochs and does not show a P-Cygni absorption trough. The blue continuum present in the ﬁrst follow-up spectrum from 2014

2http://www.astronomy.ohio-state.edu/MODS/Software/modsIDL/

15 January 29 becomes progressively weaker over time, with the spectrum from 2014 April 29 showing only slight emission above the host at wavelengths shorter than ∼ 4000 A.˚ While the latest spectrum from 2014 June 6 appears to show more blue continuum emission than the previous epoch, the ﬂux calibration and host subtraction for this spectrum are uncertain, and the corresponding Swift photometry indicates the UV emission of the source should be fading. The broad Hα emission feature becomes stronger relative to the continuum (higher equivalent width) after the initial spectrum and continues to show strong emission in all later epochs. Other broad emission features can be seen as well, including a He II 4686 line which has become stronger in equivalent width relative to the Balmer lines in the latest spectrum. We further analyze the features of these spectra in comparison to SNe, AGN, and TDEs in §2.2.3.

2.2. Analysis

2.2.1. Light Curve Analysis

After correcting both the host and transient fluxes for Galactic extinction, we produced host-subtracted light curves for all 9 photometric filters. From these data we calculated peak absolute magnitudes and decline rates for all Swift filters, which are reported in Table 2.2. Comparison with luminous SNe SN 2008es (Miller et al. 2009) and SN 2009kf (Botticella et al. 2010), both of which had absolute V -band magnitudes roughly equal to or greater than that of ASASSN-14ae, shows that the UV decline rates of these highly luminous SNe are much faster than what we observe for ASASSN-14ae, indicating that an SN explanation for the event is disfavored.

ASASSN-14ae’s (UV −UV ) and (UV −optical) color evolution are also atypical of hydrogen-rich SNe with broad lines. Figure 2.6 shows the full host-subtracted (UV − UV ) and (UV − U) color evolution for ASASSN-14ae in all Swift filters and Figure 2.7 compares the (UV M2 − UV W 1) and (UV M2 − U) colors to those of SN 2008es, a super-luminous SN IIL (Miller et al. 2009; Gezari et al. 2009), and SN 2012aw, a normal SN IIP (Bayless et al. 2013), which were also heavily observed with Swift. For SN 2008es we applied cross-filter K-corrections to obtain the rest-frame colors assuming a blackbody with Teff = 8000 K (Miller et al.

16 2009; Gezari et al. 2009), but we did not apply these corrections to SN 2012aw or ASASSN-14ae as they are much lower redshift. ASASSN-14ae shows almost no change in (UV M2 − UV W 1) color and became only slightly redder in (UV M2 − U) during the ∼ 80 days shown in the Figure. In contrast, SN 2008es became redder in both colors over time, while SN 2012aw became signiﬁcantly redder in both colors over the ﬁrst ∼ 20 days after discovery and then remained roughly constant in later epochs. ASASSN-14ae looks like neither of these, and all of its (UV − UV ) and (UV − U) colors show little change over the time shown in Figure 2.6, implying the most likely SN types that could produce the observed spectra of ASASSN-14ae are unlikely to be the sources of the transient.

2.2.2. SED Analysis

Using the host-subtracted fluxes of ASASSN-14ae we fit the transient SEDs with blackbody curves using Markov Chain Monte Carlo (MCMC) methods. The evolution of the source’s SED along with the best-fit blackbody curves are shown in Figure 2.8. At early epochs, the blackbody fit is not able to replicate the apparent excess in the UV M2 (2246 A)˚ filter. This excess is not created by the extinction correction and corresponds to no obvious emission line. Using the best-fit blackbody curves we estimate the temperature and luminosity evolution of ASASSN-14ae. The derived estimates, along with 90% confidence errors and the χ2 values of the best-fit blackbody curve, are given in Table 2.3 and shown in Figures 2.9 and 2.10. When there were Swift observations the temperature was estimated as a free parameter. When the UV data were not available, the temperature was constrained by a prior shown by the solid line in Figure 2.9 that roughly tracks the epochs with UV data. There are some degeneracies between the temperature and luminosity because much of the luminosity is farther in the UV than our data cover for these temperatures, resulting in relatively large uncertainties in some cases. In general, the temperature of the source falls from T ∼ 20, 000 K to T ∼ 15, 000 K during the first ∼ 20 days of the outburst, then rises again to T ∼ 20, 000 K over the next ∼ 50 days, and then remains roughly constant for the rest of the period shown. Conversely, the luminosity fades steadily over the 150-day period shown. Integrating over the luminosity curve using the epochs with directly estimated temperatures gives a value

17 of E ≃ 1.7 × 1050 ergs for the total energy radiated by ASASSN-14ae during this −3 −1 time. This only requires accretion of ∆M ∼ 10 η0.1 M⊙ of mass, where η =0.1η0.1 is the radiative eﬃciency.

After the first ∼ 10 days, the luminosity evolution is well fit as an exponential −t/t0 L ∝ e with t0 ≃ 39 days as shown in Figure 2.10. This differs from most TDE models where the luminosity evolution is described as a power law t−x with x ≃ 5/12 − 5/3 (e.g., Strubbe & Quataert 2009; Lodato & Rossi 2011). However, this temperature and luminosity behavior would be highly unusual for an SN, which typically exhibit a temperature that drops considerably within days of the explosion along with either a relatively constant luminosity (Type IIP; e.g., Botticella et al. 2010) or a declining luminosity (Type IIn, IIL, Ic; e.g., Miller et al. 2009; Inserra et al. 2013; Graham et al. 2014).

While it is unlikely we are seeing direct emission from a thin disk, we can model the data using the surface brightness profile of a thin disk (Shakura & Sunyaev 1973). We make the disk infinite, where the location of the inner edge is unimportant given our wavelength coverage. Adding an outer edge could be used to make the profile rise more steeply towards shorter wavelengths. Figure 2.11 shows the implied luminosity of the disk in Eddington units, where the estimate of L/LEdd depends on 6 the black hole mass MBH = 10 MBH6M⊙, the disk inclination factor cos i and the radiative efficiency η =0.1η0.1 in the sense that raising the black hole mass, making the disk inclined to the line of site or lowering the radiative efficiency will reduce the observed luminosity relative to the Eddington limit. In general, the SED of a thin disk fits the data significantly worse than a blackbody. Like the bolometric luminosity, the estimated thin disk luminosity drops exponentially with time rather −2 −3/2 than as a power law, following a plateau at L/LEdd ≃ 10η0.1MBH6 cos i for the 6.5 first ∼ 20 days. Raising the black hole mass scale to MBH ∼ 10 M⊙ would allow

L/LEdd ∼ 1 at peak. The diﬀerences compared to the Eddington limit shown in Figure 2.10 arise from the enormous increase in the unobserved hard UV emission which we discuss in §2.2.3 using the broad emission lines.

18 2.2.3. Spectroscopic Analysis

The spectra of ASASSN-14ae show broad Balmer lines in emission with a blue continuum, including an important contribution from the host galaxy (especially in the red) at late times (see Figure 2.4). We subtracted the SDSS host galaxy spectrum from ASASSN-14ae spectra in order to compare to other objects and for analyzing the spectral line proﬁles. Figures 2.12 and 2.13 show a comparison of ASASSN-14ae host-subtracted spectra at two diﬀerent epochs after discovery with the spectra of SNe II (SN 2010jl, ASASSN-13co, and SN 2008es) and a broad-line

AGN (SDSS J1540-0205). SN 2010jl is a luminous SN IIn (MV ≃ −20 mag, Stoll et al. 2011), SN 2008es is a super-luminous SN IIL (MV ≃ −22 mag, Miller et al.

2009), and ASASSN-13co is a luminous SN IIP (MV ≃ −18 mag, Holoien et al. 2016b). At the earliest epoch (Figure 2.12), the spectrum of ASASSN-14ae is similar to ASASSN-13co and SN 2008es, dominated by a blue continuum. At the later epoch (Figure 2.13), the spectrum of ASASSN-14ae stays blue, but the spectra of the SN II ASASSN-13co and SN 2008es become significantly redder, consistent with the comparison in color evolution illustrated in Figure 2.7. The spectral lines also show −1 differences. In ASASSN-14ae the Hα is broad (FWHM ∼> 8, 000 km s ) at all times and does not show the P-Cygni profile that is characteristic of SNe II with broad lines. The spectrum of ASASSN-14ae is quite different from the Type IIn SN 2010jl (Stoll et al. 2011; Zhang et al. 2012) at early and late times, both in line profiles −1 (SNe IIn have narrower lines with FWHM ∼< 5, 000 km s ) and continuum shape. The spectrum of the low-ionization broad-line AGN SDSS J1540-0205 (Strateva et al. 2003) does not resemble the spectrum of ASASSN-14ae in the earlier epoch but shows interesting similarities with the spectrum of ASASSN-14ae at 50 days after discovery. In particular, it has a complex Hα line profile, which is thought to be produced by emission from the accretion disk (Strateva et al. 2003).

In Figure 2.14 we show the evolution of the Hα line profile of ASASSN-14ae as a function of time. In the first epoch, 4 days after discovery, the line can −1 be well-fit with a Gaussian profile centered at vpeak ∼ −3, 000 km s and with −1 41 −1 FWHM ≃ 17, 000 km s and integrated luminosity LHα ≃ 2.7 × 10 erg s . However, the line peak evolves to the red and the shape becomes significantly

19 asymmetric in later epochs between ∼ 30 − 50 days after discovery. The peak of −1 the profile in these epochs is at vpeak ∼ 3, 000 − 4, 000 km s and the blue/red wing of the line reaches ∼ −15, 000/ + 20, 000 km s−1, showing a strong red asymmetry. At 70 days after discovery, the profile again becomes more symmetric −1 and can be relatively well-fit using a gaussian with vpeak ∼ +1, 400 km s , −1 41 −1 FWHM ≃ 10, 000 km s and integrated luminosity LHα ≃ 2.1 × 10 erg s . In the spectrum from 2014 April 29, not shown in Figure 2.14, the Hα line has FWHM ≃ 8, 000 km s−1 and its integrated luminosity has only decreased by a 41 −1 factor of three since the first epoch, to LHα ≃ 1.2 × 10 erg s . In the April 29 spectrum we also detect broad He II 4686 with FWHM ≃ 6, 000 km s−1 and 40 −1 LHeII ≃ 3 × 10 erg s . In the latest spectrum from 2014 June 6, this He II line has become stronger relative to the Balmer lines, but has the same FWHM.

In summary, the spectra of ASASSN-14ae seem to be inconsistent with the spectra of SNe II and the Hα emission line proﬁle shows strong evolution during the event. Compared to the spectra of TDEs in the literature, the most similar to ASASSN-14ae is SDSS TDE2 (van Velzen et al. 2011), which showed a broad Hα line with FWHM ≃ 8, 000 km s−1. The spectra of PS1-10jh (Gezari et al. 2012) showed a He II 4686 line in emission with FWHM ≃ 9, 000 km s−1 and the spectra of PS1-11af (Chornock et al. 2014) did not show any emission lines. The recent paper by Arcavi et al. (2014) presents spectra of multiple TDE candidates and shows that their spectra fall on a continuum, with some events being more He-rich and others being more H-rich. The spectra of ASASSN-14ae resembles the other TDE candidates, and with both strong He and H emission, it appears to fall in the middle of the proposed continuum of spectral properties.

If we assume that the Hα and He II emission are driven by photoionization and recombination, we can gain some insight into the hard UV continuum and the physical conditions of the line emitting region. In particular, if αB and αl are the case B recombination and line emission rate coeﬃcients, and Ei and El are the energies of the ionization edge and the line, then we can estimate the luminosity 41 at the ionization edge as Li = Ll(αB/αl)(Ei/El), which for LHα ≃ 2 × 10 and 40 −1 42 41 −1 LHeII ≃ 3 × 10 erg s implies ionizing luminosities of 3 × 10 and 2 × 10 erg s ,

20 respectively, as shown in Figure 2.15. If we compare these estimates, we see that the SED probably requires some additional hard UV emission beyond that expected from the blackbody fits, but definitely has a sharp cutoff at wavelengths only somewhat shorter than the blackbody predictions. This assumes an emission line gas covering fraction ǫ ≃ 1 near unity, and these ionizing luminosities can be shifted upwards as ǫ−1. However, the covering fractions for H and He are unlikely to differ enormously, so the SED likely must fall towards shorter wavelengths independent of the exact value of ǫ. This is also consistent with the X-ray flux limit from §2.1.2.

For comparison, Figure 2.15 shows the SED of a thin disk, raising the black 6.5 hole mass to MBH = 10 M⊙ so that an L/LEdd ≃ 1 disk with efficiency η = 0.1 is consistent with the observed UV emission near MJD 56729 (as compared with Figure 2.11). The inner regions of the disk are much hotter than the blackbody, so the SED continues to rise into the hard UV, producing far more ionizing flux than is required. This is true even when we add an inner edge at Rin = 3rg where 2 rg = GMBH /c is the gravitational radius of the black hole. Thus, while there must be some excess hard emission compared to a blackbody, it is likely less than for a thin disk. Note that the effects of the inner disk edge only affect the very hard UV and X-ray emission, which is why we could ignore the inner edge in the SED models from §2.2.2. The different shape of the thin disk model where we have the optical and near-UV SED also shows why the blackbody models are a better fit to the directly observed SEDs. Extrapolating the SED following the thin disk model would increase the total energy budget and accreted mass by roughly an order of magnitude.

We can also estimate the gas mass associated with the line emission if we assume that the line widths are related to orbital velocities. The emission radius for a velocity of v = 5000v5 km/s is of order

14 −2 rα ≃ 5 × 10 MBH6v5 cm (2.1) and the Hα luminosity is of order

4π L ≃ r3 n2α E (2.2) Hα 3 α l l 21 where αl and El are the volumetric rate and line energy and we assume a spherical geometry. This implies a characteristic number density of order

10 1/2 3 −3/2 −3 n ≃ 3 × 10 Lα41v5MBH6 cm (2.3) which implies that the line emission should almost instantaneously track the hard UV emission because the recombination times are short. The mass associated with the emission line region is then

1/2 3/2 −3 Mα ≃ 0.016Lα41MBH6v5 M⊙ (2.4) which is an order of magnitude larger than the amount of accreted mass needed to power the transient.

2.3. Discussion

The transient ASASSN-14ae, discovered by ASAS-SN on 2014 January 25, had a peak absolute V -band magnitude of MV ∼ −19.5 and position consistent to within 0.09 ± 0.14 arcseconds of the center of SDSS J110840.11+340552.2. However, it does not appear to be consistent with either an SN or a normal AGN outburst. Its colors remain blue over 140 days since detection, rather than showing the rapid reddening seen in super-luminous SNe with similar absolute magnitudes, and ASASSN-14ae’s temperature has remained relatively constant at T ∼ 20, 000 K for the duration of the outburst while declining steadily in luminosity at a rate best ﬁt by an exponential decay curve, behavior which is inconsistent with nearly all SNe. Finally, spectra of ASASSN-14ae show a strong blue continuum and broad emission features, including Balmer lines and a He II line, and its spectral evolution does not appear to match either those of SNe or AGN. While highly unusual AGN activity or a strange SN cannot be ruled out completely, the observational characteristics of ASASSN-14ae disfavor both of these scenarios.

Archival photometry, spectroscopy, and SED ﬁtting indicate that SDSS J110840.11+340552.2 appears to be an early-type galaxy with a generally intermediate-to-old-aged stellar population but some signs of recent star formation. While the recent star formation indicates that the galaxy could host a core-collapse SN, the SN explanation is disfavored by photometric and spectroscopic observations,

22 as previously mentioned. SDSS J110840.11+340552.2 shows spectral emission features indicating only a weak AGN host at best, and its mid-IR colors from WISE are inconsistent with signiﬁcant AGN activity, further disfavoring normal AGN activity as an explanation for ASASSN-14ae.

Conversely, many of the observed properties of ASASSN-14ae are consistent with previously discovered TDE candidates. The blue colors, slow decline rate, and color evolution have been seen in many TDE candidates, and these transients are predicted to show a largely constant temperature and steadily declining luminosity curve. ASASSN-14ae’s spectra also do not appear to be highly unusual for TDEs, and in fact are a very close match to spectra of the SDSS candidate TDE 2 (van Velzen et al. 2011) and the PTF candidate PTF09djl (Arcavi et al. 2014). With both strong He and H emission lines, ASASSN-14ae appears to fall in the middle of the He-rich-to-H-rich continuum proposed by Arcavi et al. (2014).

Thus, we conclude that ASASSN-14ae was most likley a TDE. At the time of discovery, ASASSN-14ae was the lowest redshift candidate TDE discovered at optical or UV wavelengths, and it continued to emit well above host galaxy levels in the UV over 140 days since discovery. In the most recent spectra, the optical continuum is again dominated by the host, but with a prominent, broad Hα line and other, weaker Balmer and He II lines.

−3 The amount of mass associated with the event is small, roughly 10 M⊙ of −2 accreted material is suﬃcient to power the transient, and ∼ 10 M⊙ is associated with the line emission region. This suggests that this event is likely powered by tidally stripping the envelope of a giant rather than by the complete disruption of a main-sequence star, as described in MacLeod, Guillochon, & Ramirez-Ruiz (2012), similar to the case seen with black hole candidate ESO 243-49 HLX-1 (Webb et al. 2014). The duration of a TDE can be truncated by putting the star on an orbit bound to the black hole (Hayasaki, Stone & LoebHayasaki et al. 2013), but this necessarily implies that a larger fraction of the stellar mass is on bound orbits and so should enhance the total energy release. On the other hand, disruptions of giants on parabolic orbits have the same asymptotic t−5/3 power law for the rate of return of material to pericenter, but they have a far higher peak. At its simplest, a constant

23 −5/3 4/3 density spherical star has a return rate proportional to P (1 − (P/P0) ) where −5/3 P0 is the period at the surface, while a shell has a rate simply proportional to P for P >P0 in both cases. As a result, disruptions of giant envelopes should show both faster rises and declines.

We continued to observe ASASSN-14ae for roughly 750 days after discovery, allowing us to study the optical evolution of the ﬂare and the transition to a host dominated state with exceptionally high precision. Our late-time observations were presented in Brown et al. (2016). We measured weak Hα emission 300 days after 39 −1 discovery, with LHα ≃ 4 × 10 ergs s , and measured the most stringent upper 39 −1 limit to-date on the Hα luminosity ∼ 750 days after discovery, LHα ∼< 10 ergs s . The observations show that optical emission from a TDE can vanish on timescales as short as 1 year, which has implications for the both the spectroscopic detection of TDEs at late times and for the nature of TDE host galaxies. We would not have been able to detect emission from ASASSN-14ae so late or place such strong limits on its emission if it were not so nearby, which demonstrates why ASAS-SN is an ideal tool for ﬁnding the TDEs that are best-suited for detailed study.

24 Fig. 2.1.— Discovery image of ASASSN-14ae. The top-left panel shows the ASAS- SN V -band reference image and the top-right panel shows the ASAS-SN subtracted image from 2014 January 25. The bottom-left panel shows the archival SDSS g-band image of the host galaxy and the bottom-right panel shows an LT 2-m g-band image from 2014 February 08. The dates of the observations are listed in each panel, and the lower panels show a smaller ﬁeld of view, indicated by the red box in the top-left panel. The red circles have radii of 5′′.0 and are centered on the host position.

25 UVW2−2.0 u+0.9 r+4.2 13 UVM2−1.2 B+1.2 i+5.3 UVW1−0.5 g+2.1 z+6.3 U+0.3 V+3.1 14

constant 18 +

19 mag

23 0 10 20 30 40 50 60 70 80 90 100110120130140150

MJD − 56682.5 (days)

Fig. 2.2.— Light curves of ASASSN-14ae starting at discovery (MJD = 56682.5) and spanning 146 days. Follow-up data obtained from Swift, LCO, and the LT 2-m are shown as circles. 3-sigma upper limits are shown as triangles for cases where the source is not detected. All magnitudes are shown in the Vega system. The data are not corrected for extinction and error bars are shown for all points. Host galaxy magnitudes measured by SDSS in a 5′′.0 aperture for ugriz and synthesized from our host SED model for the Swift UVOT bands are shown as stars at −5 days. Dates of spectroscopic follow-up are indicated with vertical bars at the bottom of the ﬁgure with colors matching the corresponding spectra in Figures 2.4 and 2.5. Table A.1 contains all the follow-up photometric data.

26 Pre-Outburst 2.5 Days 15.5 Days 52 Days 10−11 81 Days 1010 146 Days ) 1 − s 2 ) − ⊙ (L λ

−12 λL (erg cm 10 9 λ 10 λf

10−13 2000 4000 6000 800010000

Rest Wavelength (A)˚

Fig. 2.3.— Observed spectral energy distribution of ASASSN-14ae and its host galaxy. The colored squares show the SED of ASASSN-14ae at the different epochs noted in the legend (listed as days since discovery). The black circles show archival SDSS ugriz data and the black line shows the best-fit host galaxy SED from Fast. All fluxes have been corrected for Galactic extinction and all data points include error bars, although they can be smaller than the data points.

27 Fig. 2.4.— Spectral time-sequence of ASASSN-14ae during the outburst. Each spectrum shows the UT date it was obtained. Also plotted is the archival SDSS host spectrum, in black. Absorption features from the host galaxy are identiﬁed with black dotted lines. The transient spectra continue to show prominent broad Hα emission at all epochs, as well as other Balmer lines.

28 Fig. 2.5.— Host-galaxy-subtracted spectral time-sequence of ASASSN-14ae. Each spectrum shows the UT date it was obtained. Prominent emission features are identiﬁed with black dotted lines. The transient spectra show many broad emission features in all epochs, and blue continuum emission is still present at wavelengths shorter than ∼ 4000 A˚ in the latest spectrum from 2014 June 6. In later epochs, the He II 4686 line has become stronger relative to the Balmer lines.

29 4

Color + Offset −2

−4

UVW2-UVM2+3.0 UVM2-UVW1-0.6 UVW2-UVW1+1.9 UVM2-U-1.5 UVW2-U+1.0 UVW1-U-3.0 0 20 40 60 80 100 120 140

Days Since Discovery + Offset

Fig. 2.6.— (UV − UV ) and (UV − U) color evolution of ASASSN-14ae for all Swift UV bands. All fluxes used to calculate the colors shown were corrected for Galactic extinction and host-subtracted. (UV W 2 − X) colors are shown as circles colored different shades of purple, (UV M2 − X) colors are shown as squares colored different shades of blue, and (UV W 1 − U) is shown as diamonds and colored green. Each color term is offset in magnitude by a constant indicated in the legend and offset in epoch by 1 day from the term above it in order to make the plots easier to read. Horizontal dashed lines are centered on the average value of the color term plotted in the same color and are shown to aid the eye in seeing the general shape of the curves. ASASSN-14ae becomes slightly bluer in (UV W 2 − X) colors and slightly redder in (UV W 1 − U), but all terms show only slight evolution over the time shown.

30 ASASSN-14ae 3 SN 2008es SN 2012aw 2

0 UVM2-UVW1

−1

4 3 2 1

UVM2-U 0 −1 −2 0 20 40 60 80 100 120 140

Days Since Discovery

Fig. 2.7.— Comparison of (UV M2 − UV W 1) (top panel) and (UV M2 − U) (bottom panel) color evolution between ASASSN-14ae (blue circles); SN 2008es, a superluminous SN IIL (Gezari et al. 2009, green squares); and SN 2012aw, an SN IIP (Bayless et al. 2013, red diamonds). K-correction has been applied to the photometry for SN 2008es. ASASSN-14ae shows little evolution in either color while SN 2008es becomes redder in both colors and SN 2012aw becomes signiﬁcantly redder over the ﬁrst ∼ 20 days after detection and remains roughly constant thereafter.

31 2.5 Days 15.5 Days 52 Days 57 Days 10 10 61.5 Days 73 Days 81 Days 87.5 Days ) ⊙

(L 9

λ 10 λL

108

2000 4000 6000 800010000

Rest Wavelength (A)˚

Fig. 2.8.— Evolution of the SED of ASASSN-14ae (shown in different colors) along with the best-fitting blackbody models for each epoch. Only epochs with both Swift and ground data taken within 0.5 days of each other and only data points with fsub/fhost ≥ 0.3 are shown. All data points have been extinction-corrected and include error bars, although they can be smaller than the data points. At early epochs, the blackbody fits are not able to replicate the apparent excess in the Swift UV M2 band.

32 4.4

4.3

4.2

4.1

4 0 50 100 150

Fig. 2.9.— Evolution of ASASSN-14ae’s blackbody temperature with temperatures fit with a prior (open points) and without a prior (filled points). The horizontal lines show our temperature prior, with the solid line showing our central temperature prior and the dashed lines showing the 1σ spread in the prior. The temperature of the source falls from T ∼ 20, 000 K to T ∼ 15, 000 K during the first ∼ 20 days of the outburst, then rises again to T ∼ 20, 000 K over the next ∼ 50 days before remaining roughly constant for the rest of the period shown.

33 11

8 0 50 100 150

Fig. 2.10.— Evolution of ASASSN-14ae’s luminosity over time. Dashed lines show popular power law fits for TDE luminosity curves L ∝ t−x (e.g., Strubbe & Quataert 2009; Lodato & Rossi 2011) while the diagonal solid line shows an exponential fit. 6 The solid horizontal line shows the Eddington luminosity for a M = 10 M⊙ black hole. The exponential model appears to fit the luminosity curve of ASASSN-14ae better than any of the power law fits typically used for TDEs.

34 1

-1

-2

-3 0 50 100 150

Fig. 2.11.— Implied luminosity in Eddington units of ASASSN-14ae using a thin disk 6 model. The estimated L/LEdd depends on the black hole mass MBH = 10 MBH6 M⊙, the disk inclination factor cos i and the radiative efficiency η = 0.1η0.1. Raising the 6.5 black hole mass scale to MBH ∼ 10 M⊙ would produce L/LEdd ∼ 1 at peak. In general, the SED of the thin disk fits the data significantly worse than a blackbody (see Figure 2.15).

35 Fig. 2.12.— Comparison of the early-phase host-subtracted spectra of ASASSN-14ae with the spectra of the Type IIn SN 2010jl (Stoll et al. 2011; Zhang et al. 2012), the Type IIL SN 2008es (Miller et al. 2009), the Type IIP ASASSN-13co (Holoien et al. 2016b), and the broad-line AGN SDSS J1540-0205 (Strateva et al. 2003). The days with respect to maximum light (SN 2008es, SN 2010jl) or discovery (ASASSN-14ae, ASASSN-13co) are shown in parenthesis, next to the names of the transients. The same spectrum of SDSS J1540-0205 is shown in both this ﬁgure and Figure 2.13, as AGN do not have early and late phases like SNe and TDEs.

36 Fig. 2.13.— Comparison of the late-phase host-subtracted spectra of ASASSN-14ae with the spectra of the Type IIn SN 2010jl (Stoll et al. 2011; Zhang et al. 2012), the Type IIL SN 2008es (Miller et al. 2009), the Type IIP ASASSN-13co (Holoien et al. 2016b), and the broad-line AGN SDSS J1540-0205 (Strateva et al. 2003). The days with respect to maximum light (SN 2008es, SN 2010jl) or discovery (ASASSN-14ae, ASASSN-13co) are shown in parenthesis, next to the names of the transients. The same spectrum of SDSS J1540-0205 is shown in both this ﬁgure and Figure 2.12, as AGN do not have early and late phases like SNe and TDEs.

37 Fig. 2.14.— Evolution of the Hα line proﬁle of ASASSN-14ae as a function of time. We have subtracted the host galaxy spectrum and a low-order continuum deﬁned locally around the line. The days since discovery are shown in the top-right part of 41 −1 each panel. The integrated luminosity of the Hα line is LHα ≃ 2.7 × 10 erg s at 41 −1 41 −1 four days, LHα ≃ 3.3 × 10 erg s at 30 days, LHα ≃ 3.2 × 10 erg s at 51 days, 41 −1 and LHα ≃ 2.1 × 10 erg s at 70 days.

38 11

7 100 1000

6.5 Fig. 2.15.— SED of a thin disk with the black hole mass raised to MBH = 10 M⊙ so that a disk radiating at the Eddington luminosity is consistent with the observed UV emission from MJD 56729 without an inner edge (straight red dashed line), and with an inner edge at Rin =3rg (curved red dashed line). Both models rise into the hard UV, producing far more ionizing ﬂux than is required to produce the observed Hα and He II emission (unﬁlled boxes). The X-ray limit shown is based on Swift XRT data collected through 2014 April 24; including later data would make this limit tighter, as discussed in §2.1.2. These estimates of the ionizing luminosity can be shifted to higher luminosities as the inverse of the covering fraction, but H and He probably have to be shifted by similar amounts, which would imply that the spectrum must still be falling towards shorter wavelengths.

39 Filter Magnitude Uncertainty

u 19.16 0.03 g 17.60 0.02 r 16.94 0.02 i 16.65 0.02 z 16.45 0.02 J 15.34 0.05 H 14.73 0.10

KS 14.34 0.10

These are SDSS (ugriz) and 2MASS ′′ (JHKS) 5.0 radius aperture magnitudes

Table 2.1. Photometry of the Host Galaxy of ASASSN-14ae.

Absolute Magnitude Decline Rate Decline Rate Filter Magnitude Uncertainty (mag/100 days) Uncertainty

V −19.5 0.20 1.9 0.30 B −19.4 0.07 3.6 0.36 U −19.8 0.05 3.5 0.30 UV W 1 −19.8 0.04 4.0 0.30 UV M2 −20.0 0.04 3.3 0.16 UV W 2 −19.7 0.04 3.6 0.13

Table 2.2. Peak Absolute Magnitudes and Estimated Decline Rates of ASASSN-14ae in Swift Filters.

40 Luminosity Temperature Radius 2 9 4 14 MJD Best-Fit χ (10 L⊙) (10 K) (10 cm)

56684.6 35.5 21.6 ± 1.3 2.2 ± 0.1 7.0 ± 0.3 56697.4 37.2 10.7 ± 0.4 1.6 ± 0.0 9.3 ± 0.5 56698.5 27.0 9.3 ± 0.8 1.5 ± 0.1 10.4 ± 0.8 56728.6 9.8 3.1 ± 0.2 1.6 ± 0.1 5.3 ± 0.6 56734.2 15.0 2.9 ± 0.2 1.6 ± 0.1 4.8 ± 0.6 56739.6 17.2 2.6 ± 0.2 1.9 ± 0.2 3.4 ± 0.5 56744.1 1.7 2.3 ± 0.2 1.9 ± 0.2 3.1 ± 0.4 56755.2 32.1 1.7 ± 0.2 1.7 ± 0.2 3.1 ± 0.5 56763.2 9.7 1.4 ± 0.4 2.1 ± 0.5 1.9 ± 0.7 56770.3 18.2 1.5 ± 0.4 2.2 ± 0.4 1.9 ± 0.4 56794.8 28.4 0.8 ± 0.1 2.0 ± 0.2 1.7 ± 0.3

Table 2.3. Blackbody Evolution of ASASSN-14ae.

41 Chapter 3: Six Months of Multi-Wavelength Follow-up of the Tidal Disruption Candidate ASASSN-14li and Implied TDE Rates from ASAS-SN

In this chapter, I describe the discovery and follow-up observations of ASASSN-14li, the second TDE discovered by ASAS-SN and the most well-studied TDE ever found.

The ASAS-SN transient source detection pipeline was triggered on 2014 November 22, detecting a new source with V = 16.5 ± 0.1 mag (Jose et al. 2014). The object was also detected on 2014 November 11 at V = 15.8 ± 0.1 mag, but is not detected (V ∼> 17 mag) in data obtained on 2014 July 13 and before. Unfortunately, no data were obtained between 2014 July 13 and 2014 November 11 as the galaxy was behind the Sun.

A search at the object’s position (J2000 RA/Dec = 12:48:15.23/+17:46:26.22) in the SDSS DR 9 (Ahn et al. 2012) catalog revealed the source of the outburst to be the galaxy PGC 043234 (VII Zw 211) at redshift z = 0.0206, corresponding −1 −1 to a luminosity distance of d = 90.3 Mpc (H0 = 73 km s Mpc , ΩM = 0.27,

ΩΛ =0.73), and that the ASAS-SN source position was consistent with the center of the host galaxy. Follow-up images obtained on 2014 November 28 with the LCO1-m telescope at McDonald Observatory (Brown et al. 2013) and on 2014 November 30 with the Swift UVOT (Roming et al. 2005) conﬁrmed the detection of the transient.

This chapter is adapted from “Six Months of Multi-Wavelength Follow-up of the Tidal Disruption

Candidate ASASSN-14li and Implied TDE Rates from ASAS-SN”, T. W.-S. Holoien, et al. MNRAS,

455, 2918, (2016).

42 In order to constrain any offset between the source of the outburst and the nucleus of the host galaxy we first astrometrically aligned an image of the transient taken with the 2-m LT (Steele et al. 2004) with the archival SDSS image of the host galaxy. From this aligned image, we measure an offset of 0.43 ± 0.52 pixels (0.17 ± 0.21 arcseconds, or 74.4 ± 91.9 parsecs) between the position of the brightest pixel in the host galaxy in the LT image and the position of the brightest pixel in the SDSS image. This offset is consistent with the source of the outburst being the nucleus of the host galaxy, which provides support for a TDE interpretation of the event. Figure 3.1 shows the ASAS-SN V -band reference image of the host galaxy and the ASAS-SN V -band subtraction image from the discovery epoch as well as archival SDSS and post-discovery LT g-band images.

The archival SDSS spectrum of PGC 043234 shows little evidence of strong AGN activity. A follow-up spectrum of the nuclear region of the host obtained on 2014 November 30 with the SuperNova Integral Field Spectrograph (SNIFS; Lantz et al. 2004) mounted on the University of Hawaii 2.2-m telescope showed a broad Hα emission feature at the redshift of the host with a FWHM ≃ 9000 km s−1 and increased emission at bluer wavelengths. In addition to this follow-up spectrum, follow-up photometry of the source obtained on 2014 November 30 with the Swift XRT (Burrows et al. 2005) and UVOT showed strong soft X-ray emission and ultraviolet emission from a location consistent with the host nucleus. Given these observations, we determined that ASASSN-14li was a potential tidal disruption event, and began an extensive follow-up campaign in order to characterize the transient.

In §3.1 I describe pre-outburst data, including photometry and spectroscopy, of the host galaxy as well as new observations obtained of the transient during our follow-up campaign. In §3.2 I analyze these data to model the transient’s luminosity and temperature evolution and compare the properties of ASASSN-14li to those of other TDE candidates in literature. In §3.3, I use the TDE discoveries by ASAS-SN to estimate the rate of these transients in the nearby universe. Finally, in §3.4 I discuss our ﬁndings and present the results of late-time observations obtained roughly 2 years after discovery.

43 3.1. Observations and Survey Data

In this section I summarize the available archival survey data of the transient host galaxy PGC 043234 as well as our new photometric and spectroscopic observations of ASASSN-14li.

3.1.1. Archival Data

We retrieved archival reduced ugriz images of PGC 043234 from SDSS DR9 and measured the ﬂuxes in a 5′′.0 aperture radius. This aperture radius was also used to measure the source ﬂux in follow-up data, and was chosen to match the Swift PSF.

We also obtained near-IR JHKs images from 2MASS (Skrutskie et al. 2006) and measured 5′′.0 aperture magnitudes using the same procedure. Finally, we obtained a near-UV image of the host galaxy from the Galaxy Evolution Explorer (GALEX) Data Release 7 and measured the magnitude of the host using PSF photometry. The measured ﬂuxes from GALEX, SDSS, and 2MASS were then used for host galaxy SED modeling and for subtracting the host galaxy ﬂux from follow-up data containing the transient. We present the measured 5′′.0 aperture magnitudes from the SDSS and 2MASS images and the PSF magnitude from the GALEX image in Table 3.1.

There are no archival Spitzer, Herschel, HST, Chandra, or XMM-Newton observations of PGC 043234. Examining data from the ROSAT All-Sky Survey, we do not detect the host galaxy with a 3-sigma upper limit of 1.8 × 10−2 counts s−1, corresponding to 7.5 × 10−14 ergs s−1 cm−2, in the 0.08 − 2.9 keV band (Voges et al. 1999), which provides evidence that the galaxy does not host a strong AGN. However, the host galaxy is detected in the Very Large Array Faint Images of the Radio Sky at Twenty-cm survey (VLA FIRST; Becker, White & Helfand 1995), with 1.4 GHz ﬂux density of 2.96 ± 0.15 mJy, corresponding to a luminosity of 21 −1 L1.4GHz ∼ 2.6 × 10 W Hz . If this radio emission were caused by star formation, 9 the far-IR-radio correlation would imply a FIR luminosity of LF IR ∼ 3 × 10 L⊙ (Yun, Reddy & Condon 2001). However, detections of the host in archival WISE

(Wright et al. 2010) data gives a W 3-band magnitude of mW 3 = 12.367 ± 0.439,

44 7 corresponding to a luminosity of LW 3 ∼ 2 × 10 L⊙. This implies that the FIR luminosity would need to be roughly two orders of magnitude greater than the mid-IR (MIR) luminosity in order to reach the expected value if the radio emission was caused by star formation. In addition, if the galaxy exhibited the strong star formation implied by this radio emission, we would expect to see additional signs of star formation, such as strong [O II] 3727A˚ emission, which we do not see in the host spectrum (see Figure 3.3). Thus, the radio emission from the host is likely not related to star formation, but may instead be an indicator of nuclear activity. However, the galaxy has a mid-IR (MIR) color of (W 1 − W 2) ≃ 0.01 ± 0.04 in the WISE data, which, along with the non-detection in the X-ray data, provides evidence that any AGN activity is not strong (e.g., Assef et al. 2013).

Using Fast (Kriek et al. 2009), we fit stellar population synthesis (SPS) models ′′ to the 5.0 SDSS ugriz and 2MASS JHKs magnitudes of the host galaxy. The fit was made assuming a Cardelli, Clayton & Mathis (1988) extinction law with RV = 3.1 and Galactic extinction of AV = 0.07 mag based on Schlafly & Finkbeiner (2011), an exponentially declining star-formation history, a Salpeter IMF, and the Bruzual 2 & Charlot (2003) models. We obtained a good SPS fit (reduced χν =0.6), with the +0.1 9 +0.2 following parameters: M∗ = (2.8−0.1) × 10 M⊙, age= 1.3−0.3 Gyr, and a 1σ upper −2 −1 limit on the star formation rate of SFR ≤ 0.9 × 10 M⊙ yr . In order to estimate 9.3 the mass of the SMBH in PGC 043234, we use a bulge mass of MB ∼ 10 M⊙ based on the total mass from the Fast fit and Mendel et al. (2014) and the MB-MBH 6.7 relation from McConnell & Ma (2013), giving MBH ∼ 10 M⊙, a value very similar to that estimated for the host of ASASSN-14ae (Holoien et al. 2014). We find no evidence for any significant additional extinction in fits to the transient spectral energy distribution (SED) despite the fact that the Swift UV data, particularly the UV M2 band which lies on top of the 2200 A˚ extinction curve feature, is a powerful probe for additional dust. In the analyses event’s SED which follow, we only correct for Galactic extinction.

We also retrieved an archival spectrum of PGC 043234 from SDSS DR9. The archival spectrum shows both [O III] 5007A˚ and [N II] 6584A˚ emission with 5 5 luminosities of L[O III] ∼ 4.4 × 10 L⊙ and L[N II] ∼ 3.6 × 10 L⊙, respectively. As

45 with the radio emission, this emission is likely not related to star formation, as the host shows no Hα or [O II] in emission. Rather, this is again an indication that the host galaxy may host a low-luminosity AGN. However, the lack of X-ray emission and the MIR colors of the host from WISE imply that this nuclear activity is weak, as previously discussed.

The archival spectrum also shows Hδ absorption with a large equivalent width, indicating a change in the galaxy’s star formation history within the last ∼1 Gyr (e.g. Goto et al. 2003). The strong Hδ absorption indicates that the SED of the host is dominated by A stars, which implies that PGC 043234 may be a post-starburst galaxy (Goto et al. 2003). While some TDE hosts, such as NGC 5905, have shown evidence of nuclear star formation (Komossa & Greiner 1999), recent results from Arcavi et al. (2014) suggested that TDE candidates prefer post-starburst hosts, and PGC 043234 may follow this pattern. The relation between starbursts and TDEs will be further explored in a future paper on TDE hosts (Dong et al, in prep.).

3.1.2. New Photometric Observations

Following the discovery of the transient, we requested and were granted a series of 53 Swift XRT and UVOT target-of-opportunity (ToO) observations between 2014 November 30 and 2015 May 14. The UVOT observations were obtained in 6 ﬁlters: V (5468 A),˚ B (4392 A),˚ U (3465 A),˚ UV W 1 (2600 A),˚ UV M2 (2246 A),˚ and UV W 2 (1928 A)˚ (Poole et al. 2008). We extracted source counts from a 5′′.0 radius region and sky counts from a ∼40′′.0 radius region using the UVOT software task Uvotsource. We then used the most recent UVOT calibrations (Poole et al. 2008; Breeveld et al. 2010) to convert these count rates into magnitudes and ﬂuxes.

The XRT operated in Photon Counting mode (Hill et al. 2004) during our observations. The data from all epochs were reduced with the software tasks Xrtpipeline and Xselect. We extracted X-ray source counts and background counts using a region with a radius of 20 pixels (47′′.1) centered on the source position and a source-free region with a radius of 100 pixels (235′′.7), respectively. In all epochs of observation, we detect X-ray emission consistent with the position of the

46 transient. To convert the detected counts to ﬂuxes, we assume a power law spectrum with Γ = 2 and Galactic H I column density (Kalberla et al. 2005).

The XRT data have an average detected count rate of 0.3 counts s−1 in the 0.3−10 keV range, corresponding to a ﬂux of 3.0 × 10−11 ergs s−1 cm−2. This is roughly equivalent to a count rate of 1.2 counts s−1 in the ROSAT PSPC, indicating an increase by a factor of ∼ 60 over the ROSAT limits.

In addition to the Swift XRT and UVOT observations, we also obtained ugriz images with the LT 2-m telescope and the LCO 1-m telescopes at Siding Spring Observatory, McDonald Observatory, Sutherland, and Cerro Tololo. We measured aperture photometry using a 5′′.0 aperture radius in order to match the host galaxy and Swift measurements1. We determined photometric zero-points using several SDSS stars in the ﬁeld.

Figure 3.2 shows the X-ray, UV, and optical light curves of ASASSN-14li. The XRT flux measurements and UVOT/ugriz magnitudes are presented in Table A.2 and Table A.3, respectively. The observations cover the period from MJD 56983.6 (the epoch of discovery) through our latest epoch of observations on MJD 57161.9, spanning 178.3 days. The data are shown without extinction correction or host flux subtraction. Also shown in Figure 3.2 are the host magnitudes measured from SDSS images for ugriz and extrapolated from the host SED fit for Swift UVOT filters. Although our observations likely missed the peak of the transient’s light curve, they still show that ASASSN-14li brightened considerably with respect to the host galaxy in the UV and blue filters, with the largest increase being in the Swift UV W 2 band, where it brightened by ∆mUVW 2 ∼ −4.1. The g-band increase was significantly weaker, with ∆mg ∼ −0.4.

1We attempted to do image subtraction of the follow-up ugriz data with the SDSS archival images as templates. However, due to the lack of stars in the ﬁeld-of-view close to ASASSN-14li, the quality of the subtractions was sub-optimal and accurate measurements were not possible.

47 3.1.3. New Spectroscopic Observations

We obtained spectra of ASASSN-14li spanning 145 days between UT 2014 December 02 and UT 2015 April 14. The spectrographs used for these observations, along with wavelength range and resolution in angstroms, were: SNIFS mounted on the 2.2-m University of Hawaii telescope (3200 − 10000 A,˚ R ∼ 3 A),˚ the Dual Imaging Spectrograph (DIS) mounted on the Apache Point Observatory 3.5-m telescope (range 3500 − 9800 A,˚ R ∼ 7 A),˚ the Multi-Object Double Spectrographs (MODS; Pogge et al. 2010) mounted on the dual 8.4-m Large Binocular Telescope (LBT) on Mount Graham (3200 − 10000 A,˚ R ∼ 3 A),˚ the Modular Spectrograph (Modspec) mounted on the MDM Observatory Hiltner 2.4-m telescope (4660−6730 A,˚ R ∼ 4 A),˚ the Ohio State Multi-Object Spectrograph (OSMOS; Martini et al. 2011) mounted on the MDM Observatory Hiltner 2.4-m telescope (4200 − 6800 A,˚ R ∼ 4 A),˚ the Fast Spectrograph (FAST; Fabricant et al. 1998) mounted on the Fred L. Whipple Observatory Tillinghast 1.5-m telescope (3700 − 9000 A,˚ R ∼ 3 A),˚ and the Inamori Magellan Areal Camera and Spectrograph (IMACS; Dressler et al. 2011) mounted on the Las Campanas Observatory Magellan-Baade 6.5-m telescope (3650 − 9740 A,˚ R ∼ 8 A).˚ The spectra from MODS were reduced using a custom pipeline written in Idl2 while all other spectra were reduced using standard techniques in Iraf. Telluric corrections were applied to the spectra using the spectra of spectrophotometric standard stars observed on the same nights. We calculated synthetic r-band magnitudes and scaled each spectrum to match the r-band photometry. Figure 3.3 shows a time-sequence of selected ﬂux-calibrated follow-up spectra along with the archival SDSS spectrum of the host. Figure 3.4 shows a time-sequence of the same spectra with the host galaxy spectrum subtracted. Summary information, including dates, instruments used, and exposure times for all follow-up spectra, are listed in Table 3.2.

The key characteristics of the spectra of ASASSN-14li are a strong blue continuum, consistent with the photometric measurements, and the presence of broad Balmer and helium lines in emission, which are either absent or seen as absorption features in the host spectrum. The emission features are highly

2http://www.astronomy.ohio-state.edu/MODS/Software/modsIDL/

48 asymmetric and have widths of ∼ 10, 000 − 20, 000 km s−1 in all epochs, though they appear to grow narrower over time. The blue continuum becomes progressively weaker over time, though it still shows emission in excess of the host at wavelengths shorter than ∼ 6000 A˚ in the latest epoch, which is in agreement with the UV and optical photometry. We further analyze the features of these spectra and compare to ASASSN-14ae and other TDE candidates in §3.2.2.

3.2. Analysis

3.2.1. SED Analysis

Using the 5′′.0 aperture magnitudes measured from the archival SDSS data and synthesized from the Fast SED fit for Swift UVOT filters we performed host flux subtraction on the extinction-corrected photometric follow-up data. We then used these host-subtracted fluxes to fit the SED of ASASSN-14li with blackbody curves using MCMC methods, as was done for ASASSN-14ae in Holoien et al. (2014). As flux in redder filters was clearly dominated by host flux even in early epochs, we only use filters with effective wavelength less than 4000 A˚ (Swift U, UV W 1, UV M2, and UV W 2 and SDSS u) when fitting the SED. Unlike ASASSN-14ae, the UV data for ASASSN-14li do not appear to span the peak of the SED, resulting in a broad range of possible blackbody temperatures. If we use a very weak temperature prior (±1 dex), the median epoch with measurements at four or more wavelengths has a formal temperature uncertainty of ±0.2 dex. However, without a clear detection of the spectral peak, temperature uncertainties are likely dominated by systematic errors (e.g. host flux or deviations of the true spectral shape from a black body) rather than this estimate of the statistical errors. We therefore adopted a temperature typical of these weakly constrained fits and a strong prior of log T/K =4.55 ± 0.05 for our standard fits. The evolution of the source’s SED along with the best-fit blackbody curves are shown in Figure 3.5.

With the temperature constrained by our prior to remain roughly constant, the optical/UV luminosity of the source fades steadily over the ∼ 178 days after initial

49 discovery. The luminosity evolution is well ﬁt by an exponential curve L ∝ e−t/t0 with t0 ≃ 60 days, as shown in Figure 3.6. This diﬀers from common TDE models, where the luminosity evolution is expected to follow a power law t−x with x ≃ 5/12 − 5/3 (e.g., Strubbe & Quataert 2009; Lodato & Rossi 2011). However, the exponential luminosity evolution matches that of ASASSN-14ae (Holoien et al. 2014), meaning that this behavior is consistent with previously discovered TDE candidates. This temperature and luminosity behavior is inconsistent with what would be expected if ASASSN-14li were an SN (which typically show rapidly declining temperatures along with constant or declining luminosity (e.g., Miller et al. 2009; Botticella et al. 2010; Inserra et al. 2013; Graham et al. 2014), providing evidence that ASASSN-14li is a better match to a TDE than an SN.

The X-ray luminosity, by contrast, declines at a much slower rate than the optical/UV luminosity, and roughly 40 days after discovery the X-ray luminosity becomes the dominant source of emission. As shown in Figure 3.5, the X- ray luminosity requires a significantly higher blackbody temperature (roughly T ∼ 105 K) than the T ∼ 35, 000 K temperature that best-fits the optical/UV data. While we cannot rule out a single blackbody component with much higher temperature generating both the optical/UV and the X-ray emission, we believe it is more likely that the the X-ray emission arises from a different, hotter region of the source seen through a region of lower-than-average density, as described in Metzger & Stone (2016), or that it is non-thermal.

Integrating over the X-ray and optical/UV luminosity curves using only epochs with Swift UV data implies that ASASSN-14li radiated a total energy of E ≃ 7 × 1050 ergs over the time period covered by our follow-up data. This requires −3 −1 accretion of only ∆M ∼ 4.0 × 10 η0.1 M⊙ of mass, where η0.1 =0.1η is the radiative eﬃciency, to power the event.

3.2.2. Spectroscopic Analysis

In their analysis of TDE candidates discovered by the Palomar Transient Factory, Arcavi et al. (2014) found that all their candidates showed similar spectroscopic

50 characteristics, including a strong blue continuum, and that their candidates spanned a continuum from H-rich to He-rich spectroscopic features. The follow-up spectra of ASASSN-14li show many of the same characteristics as the TDEs that Arcavi et al. (2014) refer to as “intermediate H+He events,” because its spectra have strong He and H emission features in all epochs. In Figure 3.7 we compare the spectra of ASASSN-14li to ASASSN-14ae from Holoien et al. (2014) at similar epochs after discovery. The spectra are broadly similar, as expected, with both showing strong emission at bluer wavelengths and broad Hα emission features, but there are some notable differences. The blue continuum seems to fade somewhat more slowly in ASASSN-14li than in ASASSN-14ae, and ASASSN-14li shows a broad He II 4686 A˚ emission feature even in early epochs (a few days after discovery), whereas ASASSN-14ae only began to show this feature a few months after discovery. We note that some of these differences could stem from the fact that ASASSN-14li was further past its peak luminosity when discovered than ASASSN-14ae was (the previous non-detection of ASASSN-14ae was roughly 3 weeks prior to discovery, whereas ASASSN-14li was not observed for nearly 3 months before being discovered). However, given the range of TDE candidate properties found by Arcavi et al. (2014), it is likely that some of the differences observed between these objects are unrelated to their age.

In Figure 3.8 we show the luminosity evolution of three strong emission features (Hα, Hβ, and He II 4686 A)˚ present in the spectra of ASASSN-14li. As estimating the true error on these fluxes is difficult given their complex shape, we assume 20% errors on the emission fluxes calculated in each epoch. The three lines span a wide 41 −1 range in luminosity in early spectra, with peak values of LHα ∼ 1.1 × 10 ergs s , 40 −1 40 −1 LHβ ∼ 4.0 × 10 ergs s , and LHeII ∼ 8.5 × 10 ergs s . However, the more luminous lines also appear to decline in luminosity more quickly, such that by the time of our most recent spectra, all three lines have luminosities in the range L ∼ (0.5 − 1.5) × 1040 ergs s−1. In Figure 3.8 we also show the Hα emission expected given the measured Hβ emission assuming the emission is driven by case B recombination. Within noise, the Hα/Hβ ratio seems to be largely consistent with recombination. The measured luminosities for all three lines are given in Table 3.3.

51 We also compare the evolution of the Hα and He II 4686 A˚ line profiles of ASASSN-14li to those of ASASSN-14ae in Figures 3.9 and 3.10, spanning the period from 9 days after discovery to 145 days after discovery for ASASSN-14li and from 4 days after discovery to 132 days after discovery for ASASSN-14ae. While both objects show strong emission features, the evolution of these features shows a number of differences between the two objects. The Hα profile is fairly asymmetric and narrows over time for both objects (see Figure 3.9. However, ASASSN-14li shows a significantly narrower Hα emission feature than ASASSN-14ae in all epochs. At 9 days, the ASASSN-14li Hα feature shows a narrow peak and has blue/red wings reaching ∼ −10, 000/+10, 000 km s−1 at the base of the line, while at 4 days the ASASSN-14ae Hα feature shows a broad peak and has blue/red wings reaching ∼ −15, 000/+10, 000 km s−1. The Hα feature becomes significantly narrower for both objects by the time of the latest spectroscopic epoch, with ASASSN-14li showing blue/red wings reaching ∼ −5, 000/+5, 000 km s−1 and ASASSN-14ae showing blue/red wings reaching ∼ −5, 000/+10, 000 km s−1. The FWHM of the two lines show a similar evolutionary trend, with the Hα feature narrowing from FWHM ≃ 3, 000 km s−1 to FWHM ≃ 1, 500 km s−1 for ASASSN-14li and from FWHM ≃ 17, 000 km s−1 to FWHM ≃ 8, 000 km s−1 for ASASSN-14ae. Given the fact that “later” epochs of ASASSN-14ae seem to resemble “earlier” epochs of ASASSN-14li, the Hα evolution suggests that ASASSN-14li may have been older than ASASSN-14ae at discovery. However, the uncertainty on the age of ASASSN-14li is not so great as to suggest that all these differences are strictly due to the age of the transient.

The two TDE candidates show more obvious diﬀerences in the evolution of their He II 4686 A˚ line proﬁles. As can be seen in Figure 3.10, ASASSN-14li displays a strong He II emission feature in all epochs, while ASASSN-14ae only begins to show a similar feature in later spectra. The He II feature shows similar evolution to the Hα feature for ASASSN-14li, as it is fairly asymmetric in all epochs and narrows over time, with the 9-day spectrum showing blue/red wings reaching ∼ −10, 000/+5, 000 km s−1 and the 145-day spectrum showing blue/red wings reaching ∼ −5, 000/+3, 000 km s−1. ASASSN-14ae does not show a He II feature in the 4-day spectrum, but its latest spectrum taken at 132 days after discovery

52 shows a feature that resembles that of ASASSN-14li, with blue/red wings reaching ∼ −5, 000/+3, 000 km s−1.

Finally, we examine the evolution of the emission line widths, shown in Figure 3.11 for Hα. As the luminosity of the transient is decreasing, so too is the line width. This is the opposite of what is seen in reverberation mapping studies 2 1/2 of quasars, where estimates of the black hole mass MBH ∝ ∆v L remain roughly constant because the luminosity decreases as the line width broadens (e.g., Peterson et al. 2004; Denney et al. 2009). Physically, this is believed to result from the fact that if the luminosity drops, gas at larger distances and lower average velocities recombines, leading to an increase in the line width.

Simple estimates based on the line luminosities and light travel times imply that the densities of the ionized regions of ASASSN-14li are also high enough to make the recombination times negligible. This means it is unlikely that the narrowing of the lines is due to a ﬁnely-tuned outward density gradient allowing the higher velocity material at smaller radii to recombine faster while also making the line width shrink with time. The fast recombination times also make it diﬃcult to explain the decreasing line widths as simply being due to a continuing expansion of the Stromgren sphere even as the luminosity is decreasing.

No TDE has been caught early enough to make these measurements, but there should be temporal lags between the rise of the UV emission and the formation of the broad lines, as is seen in reverberation mapping of AGN (e.g., Peterson et al. 2004). At late times, these eﬀects are still present, but seem an unlikely explanation for the observed line width evolution, as the temporal smoothing of the changes becomes larger when the delays are long.

A final possibility is that the changes represent evolution in the density distribution of the ionized gas. While there is no reason to expect rapid, large scale gas redistributions in a normal AGN, such changes seem plausible during a TDE. This is presumably not a large scale redistribution, since the time scale for a −3 7 significant change in radius is t ∼ 50MBH7v3 years, where MBH7 = (10 M⊙)MBH −1 and v3 = (3, 000 km s )v. It would not be surprising, however, to have significant

53 evolution in the mean density at a given radius given the nature of a TDE. For example, if the line emission is dominated by a dense gas phase, but the average density of this phase is decreasing, the total line emission diminishes in proportion to the density. Given the quality of the spectra of ASASSN-14li, we do not attempt any quantitative analysis, but these questions suggest a need for higher quality spectra in future studies of TDEs.

The spectra of ASASSN-14li seem to be consistent with both those of ASASSN-14ae and those of other TDE candidates in literature, showing strong emission at bluer wavelengths that fades steadily over time and strong Balmer and helium emission features in all epochs. These similarities, as well as the fact that these spectra do not seem to resemble those of type II SNe or AGN, provide strong evidence for a TDE interpretation for ASASSN-14li.

3.3. TDE Rates

Due to the small number of candidate TDEs, the rate of stellar tidal disruptions by SMBHs is not particularly well-understood. This rate may be important, however, as it can be used to study the orbits of stars around SMBHs and the origin of relativistic TDEs that have been observed with Swift (van Velzen & Farrar 2014). Furthermore, if the rate is high enough, it may even play an important role in the growth of SMBHs (e.g., Magorrian & Tremaine 1999). Van Velzen & Farrar (2014) used four TDE detections (two from SDSS and two from Pan-STARRS) and a search +2.7 −5 −1 of SDSS Stripe 82 galaxies to estimate a TDE rate of (1.5 − 2.0)−1.3 × 10 yr per galaxy. In Holoien et al. (2014) we used estimates for the local density of black holes from Shankar (2013) and an observable volume of 3 × 107 Mpc3 for ASAS-SN to estimate that ASAS-SN would discover 0.3 − 3 TDEs per year, assuming a 50% detection efficiency. If we take our estimate of 50% detection efficiency to be true, the fact that ASAS-SN discovered 2 TDE candidates in 2014 would imply that the actual rate of tidal disruptions is actually much higher than the estimate from van Velzen & Farrar (2014), possibly as high as ∼ 10−4 yr−1 per galaxy. However, the assumption of a 50% detection efficiency is arbitrary, and as the rate estimate

54 depends strongly on this quantity, a more realistic simulation of the ASAS-SN detection eﬃciency is needed.

To make a more quantitative estimate of the TDE rate, we used SDSS galaxies, which includes the hosts of both of our TDE candidates. We selected all galaxies from SDSS DR9 with 0.01

We modeled the peak of the light curve, which will dominate any magnitude 2 2 limited detections, as MV = Vpeak +(t − t0) /t1 where Vpeak is the peak absolute magnitude, t0 is the time of peak, and t1 is the time to decay by one magnitude from the peak. For each galaxy, these were converted to apparent magnitudes −1 −1 using a quadratic ﬁt to the luminosity distance for an H0 = 70 km s Mpc ,

Ω0 = 0.3, ΩΛ = 0.7 ﬂat cosmological model combined with the Galactic extinction in each sightline. For each galaxy, Nt = 1000 trial values of t0 were drawn from one month before the start of the light curve to one month afterward for a total time span of ∆tm, and the trial source was viewed as detected if there would be two signal-to-noise ratio S/N > 7.5 detections within one week given the model and the estimated noise in the actual light curve for the center of the galaxy. This criterion detects both of our observed candidates. With these deﬁnitions, the survey time over which we would detect a TDE in any single galaxy is ∆t = ∆tmNd/Nt, where

Nd is the number of trial detections. Averaging this over all the trial galaxies gives us a mean survey time per galaxy h∆ti, leading to a rate estimate of N r = TDE . (3.1) Ngh∆ti

The results as a function of Vpeak and t1 are presented in Table 3.4. We have not included results for transients fainter than Vpeak = −18 mag because they begin to be significantly affected by the z > 0.01 redshift cutoff.

Using ASASSN-14ae (Holoien et al. 2014) and PS1-10jh (Gezari et al. 2012) as examples of “typical” TDEs, we assume that typical TDEs have Vpeak ≃ −19 to −20 and decay by one magnitude in 25 to 40 days. In Table 3.4, we see

55 that the time scale t1 has only a modest affect on the effective survey time once t1 > 10 days, while the peak magnitude has an enormous effect because it controls the effective survey volume. If we assume that TDEs are uniformly distributed over the Vpeak = −19 and −20 mag bins for t1 = 20, 30 and 40 days, we find that −5 −1 Ngh∆ti ≃ 49, 000 years implying an average TDE rate of r ≃ 4.1 × 10 yr per galaxy given NTDE = 2. The Poisson uncertainties correspond to a 90% confidence range of (2.2 − 17.0) × 10−5 yr−1 per galaxy, and these probably dominate over the systematic uncertainties. For example, the fractional shifts from taking any of the cases averaged over to yield the estimate of Ngh∆ti are significantly smaller than the Poisson uncertainties. This rate estimate is higher than the rate of +2.7 −5 −1 (1.5 − 2.0)−1.3 × 10 yr per galaxy found by van Velzen & Farrar (2014), though the two estimates are consistent given the uncertainties. If we lower the threshold to two observations with S/N > 5, the rate estimate drops by a factor of 1.5 and is closer to the estimate of van Velzen & Farrar (2014), but the predicted magnitude distribution of Type Ian SN for this threshold is somewhat fainter than is observed. If we raise the threshold to S/N > 10, the rate estimate rises by a factor of 1.4 and is closer to theoretical predictions (e.g., Stone & Metzger 2016; Metzger & Stone 2016), but the predicted magnitude distribution of Type Ian SN is somewhat brighter than is observed and the two TDEs no longer satisfy the detection criterion.

3.4. Discussion

ASASSN-14li, discovered by ASAS-SN on 2014 November 11, had a position consistent to within 0.17 ± 0.21 arcseconds of the center of PGC 043234 and a peak absolute V -band magnitude of MV ∼ −19. Follow-up observations indicate that it is not consistent with either an SN or a normal AGN outburst. Conversely, it shows many similarities with other TDE candidates discovered by optical surveys. ASASSN-14li has remained bright in the UV and blue optical filters even six months after detection, and the best-fit blackbody temperature has remained roughly constant at T ∼ 35, 000 K for the duration of the outburst while the luminosity has declined at a steady rate best fit by an exponential decay curve. Spectra of ASASSN-14li show a strong blue continuum and broad Balmer and helium emission

56 features in all epochs, and do not show the spectral evolution expected of SNe or AGN. Such features are characteristic of TDE candidates such as those discovered by ASAS-SN (ASASSN-14ae, Holoien et al. 2014) and iPTF (Arcavi et al. 2014), which leads us to the conclusion that ASASSN-14li was likely a TDE as well. Its proximity (z = 0.0206) makes it the closest TDE candidate discovered at optical wavelengths to date, and it is also the ﬁrst TDE candidate discovered at optical wavelengths to exhibit both UV/optical emission as well as associated X-ray emission.

Archival spectroscopy and photometry, as well as SED ﬁtting, indicate that the transient host galaxy PGC 043234 has undergone a change in its star formation history within the last ∼ 1 Gyr and has a stellar population dominated by A stars, implying that it is a post-starburst galaxy. Previous work by Arcavi et al. (2014) indicated that TDE candidates may prefer such hosts, and this possibility will be further explored in future work on ASAS-SN TDE host galaxies (Dong et al., in prep.). The host does not show signs of recent star formation, and while the detection of [O III] emission and radio emission may imply that it hosts a weak AGN, its mid-IR colors from WISE are inconsistent with signiﬁcant AGN activity.

ASASSN-14li has significantly more He II emission than ASASSN-14ae, so much so that the He II Stromgren sphere is likely of comparable size to the H Stromgren sphere. This is consistent with the diversity of He and H line strengths noted by Arcavi et al. (2014). The observed soft UV SEDs, line luminosities, and X-ray properties of the two ASAS-SN TDE candidates also provide a natural explanation of the diversity. ASASSN-14li has both a harder soft UV continuum and significant X-ray flux, which implies a stronger hard/ionizing UV continuum. This is then observed indirectly through the greater ratio of He to H emission in the spectra. While the unknown covering fractions mean that the observed line luminosities only set lower bounds on the hard UV continuum, their ratios likely provide reasonable estimates of the spectral slopes because it is difficult to make the H and He line emission regions enormously different.

The evolution of the line widths is also interesting, as discussed in more detail in § 3.2.2. For the “quiescent” environments of reverberation mapped AGN, decreases in luminosity are accompanied by increases in line widths (e.g. Peterson

57 et al. 2004; Denney et al. 2009). For ASASSN-14li we see the opposite: as the transient fades, the line widths become narrower. One likely explanation for this behavior is that the line width evolution is due to more rapid evolution of the density distribution closer to the black hole. We do not mean in the sense of large scale changes in the amount of mass at a given radius – for the typical velocities of the lines (1,000-3,000 km s−1), this occurs relatively slowly (years to decades) if the velocities are virialized. However, rearrangements of the gas that change the mean emission measure can occur much more rapidly, so the line evolution may provide a probe of the evolution of the “clumpiness” of the material on these velocity scales.

Both of these issues are worth exploring in more detail, but they also need higher quality spectra than we have available for ASASSN-14li. Spectra with signiﬁcantly higher signal-to-noise ratios obtained at a higher cadence are needed to closely track the evolution of the structure of the emission lines. In such a close study it will be important to bear in mind the eﬀects of light travel times (days to months), potentially with the possibility of essentially carrying out a reverberation mapping study of a TDE. This would also require a high cadence continuum light curve including the onset of the transient.

We used SDSS galaxies including the hosts of the two TDE candidates discovered by ASAS-SN to estimate the rate of tidal disruptions in galaxies with redshift 0.01

We also note that if we carry out a similar rate analysis for Type Ia SNe in ASAS-SN, we ﬁnd a rate consistent with that found by the LOSS survey (Li et al.

58 2011), suggesting that the overall analysis is reasonably robust. A detailed analysis of Type Ia rates in ASAS-SN will be carried out in a future publication by Holoien et al. (in prep.).

TDEs and Type Ia SNe have broadly similar peak magnitudes and time scales at peak, so we would naively expect the two source types to have similar selection effects and hence ratios of event numbers between surveys. For ASAS-SN, this ratio is approximately 1 TDE for every 70 Type Ian SNe (2 and 135, respectively). The ratios for PTF (Arcavi et al. 2014; Rau et al. 2009), Pan-STARRS and SDSS (Gezari et al. 2012; Chornock et al. 2014; van Velzen & Farrar 2014) are are 1 TDE for every 550, 1000, and 1050 SNe Ia, respectively.3 Unless we have been unusually lucky (for the PTF/Pan-STARRS/SDSS ratios, we would have a 1-2% probability of finding two TDEs), this comparison suggests the (testable) possibility that the completeness of TDE searches in ASAS-SN is markedly higher than in prior surveys. This offers the exciting possibility that a much larger population of TDEs can be identified in existing surveys, although the sample found by ASAS-SN will remain the most useful for detailed study due to their intrinsic brightness. It remains to be seen, however, if this will help explain the gap between observed and theoretical TDE rate estimates (see Metzger & Stone 2016).

ASASSN-14li is the brightest TDE candidate ever discovered at optical wavelengths, and it has been extensively observed for over 6 months since discovery for our initial study, resulting in an unprecedented data set spanning optical to X-ray wavelengths. In Brown et al. (2017) we presented the results of additional observations spanning roughly 600 days after discovery. The latest optical spectra show Hα emission well in excess of that seen in the SDSS host galaxy spectrum, representing the latest detection of line emission in TDE spectra and indicating that the processes powering the luminous ﬂares associated with TDEs can operate for several hundreds of days. Late-time Swift observations reveal lingering thermal UV and X-ray emission with characteristic temperatures evolving by at most a factor of ∼ 2 during the observation campaign. The late-time evolution of the X-ray, UV,

3This is based on 3, 2 and 2 TDE candidates in samples of 550, > 2000 (Jones, Riess & Scolnic

2015), and ∼ 2100 (Sako et al. 2014) Type Ian SNe.

59 and line emission is consistent with a power-law decay rate. The long-lasting nature of ASASSN-14li is in stark contrast to most optically discovered TDEs, which, like ASASSN-14ae, are typically X-ray faint and evolve on shorter timescales. Because ASASSN-14li was so long-lived, it presents new opportunities for discovering TDEs through the observation of lingering TDE emission features in potential hosts, and the complete dataset of observations now includes thousands of observations in wavelengths ranging from the hard X-ray to radio.

60 Fig. 3.1.— Discovery image of ASASSN-14li. The top-left panel shows the ASAS- SN V -band reference image and the top-right panel shows the ASAS-SN subtracted image from 2014 November 22, the date of discovery. The bottom-left panel shows the archival SDSS g-band image of the host galaxy PGC 043234 and the right panel shows an LT 2-m g-band image taken during our follow-up campaign. The dates of the observations are listed on each panel. The red circles have radii of 3′′.0 and are centered on the position of the transient measured in the LT image.

61 10 − + + XRT UVW1 0.5 B 1.2 r 4.2 −10 UVW2−2.0 U+0.3 g+2.1 i+5.3 10 11 UVM2−1.2 u+0.8 V+3.1 z+6.3

13 10−11

14 ) 2

15 10−12 16 constant + 17 mag 18 10−13 X-ray Flux (ergs /s /cm

10−14 21

22 0 20 40 60 80 100 120 140 160 180

Days Since Discovery (MJD=56983.6)

Fig. 3.2.— Light curves of ASASSN-14li beginning on the epoch of discovery (MJD= 56983.6) and spanning 178 days. Follow-up data obtained from Swift (X- ray, UV, and optical; circles), the LT 2-m (optical; squares), and the LCO 1-m telescopes (optical; diamonds) are shown as circles. All UV and optical magnitudes are shown in the Vega system (left scale), and X-ray fluxes are shown in ergs/s/cm2 (right scale). The scales are chosen so that time variability has the same meaning for both the X-ray and optical/UV data. The data are not corrected for extinction and error bars are shown for all points. Observations in the riz and Swift BV filters were terminated earlier than those in other bands, as the source had faded to be fainter than the host. 5′′.0 aperture magnitudes measured from archival SDSS images for ugriz and synthesized from our host SED model for the Swift UVOT filters are shown as stars at −5 days. Vertical bars at the bottom of the figure indicate dates of spectroscopic follow-up. Table A.3 contains all the follow-up photometric data.

62 2014/12/02 2014/12/12 4 2014/12/14 2014/12/19 2015/01/03 2015/02/16 2015/03/11 ˚ A) /

2 3 2015/04/17 SDSS Host erg/s/cm 15

− 2 10 ( λ f

4000 5000 6000 7000 8000 9000 Rest Wavelength (A)˚

Fig. 3.3.— Spectroscopic time-sequence of ASASSN-14li. Colors indicate the dates of observation. The archival SDSS spectrum of PGC 043234 is plotted in black.

63 3 Hδ 2014/12/02 2014/12/12 He II 2014/12/14 Hγ 2014/12/19 2015/01/03 Hβ 2015/02/16 2 2015/03/11 ˚ A) /

2 2015/04/17 erg/s/cm Hα 15

− 1 10 ( He I λ f

4000 5000 6000 7000 8000 Rest Wavelength (A)˚

Fig. 3.4.— Host-galaxy-subtracted spectroscopic time-sequence of ASASSN-14li, showing the same spectra used in Figure 3.3 with the host galaxy spectrum subtracted. Prominent emission lines in the transient spectrum, including Balmer lines, He I 5725A,˚ and He II 4686A,˚ are indicated by vertical dotted lines. The transient spectra show many broad emission features and blue continuum emission above that of the host at wavelengths shorter than ∼ 4000 A˚ in all epochs

64 8 Days 24 Days X-ray 10.0 33 Days 48 Days 59 Days 74 Days 9.5 82 Days

) 94 Days ⊙ 106 Days 128 Days ) (L λ 9.0 Hα 143 Days 173 Days λL log( 8.5 He II

8.0

102 103 104

Rest Wavelength (A)˚

Fig. 3.5.— Evolution of the SED of ASASSN-14li shown in different colors (points) and the corresponding best-fitting blackbody models for each epoch (lines). Only data points with subtracted flux greater than 0.3 times the host flux are shown, as data points with a smaller fractional flux were highly uncertain and contributed little to the fit. All data points have been corrected for Galactic extinction and include error bars, though in some cases the error bars can be smaller than the data points. All fits were made assuming a temperature prior of log T/K = 4.55 ± 0.05. Also shown are the early-epoch X-ray luminosity and ionizing luminosities implied by the Hα and He II 4868A˚ lines (see § 3.2.2.) No additional UV emission is required to produce the line emission, but the X-ray emission is likely non-thermal or produced in a separate, hotter region.

65 11

10.5

9.5

8.5 0 50 100 150 200

Fig. 3.6.— Evolution of ASASSN-14li’s luminosity over time. Open triangles indicate the X-ray luminosity measurements. Dashed lines show popular power law fits for TDE luminosity curves L ∝ t−x (e.g., Strubbe & Quataert 2009; Lodato & Rossi 2011) while the diagonal solid line shows an exponential fit. The Eddington luminosity for a 6 M = 10 M⊙ black hole is shown as a solid horizontal line. The luminosity evolution appears to be best-fit by the exponential model, similar to the previous ASAS-SN TDE candidate, ASASSN-14ae (Holoien et al. 2014).

66 ASASSN-14li ASASSN-14ae 8.5

8.0 +9 days

+4 days 7.5

+30 days

) + constant 7.0 λ

+30 days log(f 6.5

+85 days 6.0 +73 days

5.5 4000 5000 6000 7000 8000 9000 Rest Wavelength (A)˚

Fig. 3.7.— Comparison of the spectra of ASASSN-14li with those of ASASSN-14ae (Holoien et al. 2014) at similar epochs after discovery. ASASSN-14li is shown in blue while ASASSN-14ae is shown in red, and epochs in days-after-discovery are shown to the right of each spectrum. The spectra look broadly similar, with both objects showing a strong blue continuum that declines over time and a broad Hα emission feature in all epochs. However, the spectra of ASASSN-14li also show a strong He II 4686 A˚ emission feature in all epochs, whereas ASASSN-14ae only begins to show He II emission in later epochs.

67 Case B Hα Hα 1041 Hβ He II ) 1 − Luminosity (ergs s

1040

0 20 40 60 80 100 120 140 Days After Discovery

Fig. 3.8.— Evolution of the Hα (red squares), Hβ (green triangles), and He II 4686 A˚ (blue circles) luminosities. Error bars show 20% errors on the line ﬂuxes. The black line shows the Hα emission that would be expected from case B recombination, given the Hβ emission. The Hα emission is generally consistent with recombination.

68 1.0 9 26 4 30 0.8

0.6

0.4 ) 0.2 λ,cont f

− 0.0 λ

1.0 85 145 73 132 0.8

Normalized (f 0.6

0.4

0.2

0.0

-2 -1 0 1 2 -1 0 1 2 v(Hα)/104 (km s−1)

Fig. 3.9.— Evolution of the Hα line profile of ASASSN-14li (black) and ASASSN-14ae (red; Holoien et al. 2014). The number of days since discovery for each spectrum is shown in the upper-right corner of each panel, with colors matching the colors of the spectra. We have subtracted the host galaxy spectra and a locally defined low-order continuum around the lines from each spectrum. Both objects show an asymmetric Hα profile that narrows over time, with ASASSN-14li showing a significantly narrower profile than ASASSN-14ae in all epochs.

69 1.0 9 26 4 30 0.8

0.6

0.4

) 0.2 λ,cont

f 0.0 − λ −0.2

1.0 85 145 73 132 0.8

0.6 Normalized (f 0.4

0.2

0.0

−0.2 −2 −1 0 1 2 −1 0 1 2 v(He II)/104 (km s−1)

Fig. 3.10.— Evolution of the He II 4686 A˚ line profile of ASASSN-14li (black) and ASASSN-14ae (red; Holoien et al. 2014). (The strong line to the right of the He II 4686 A˚ line in the right panel is Hβ.) The number of days since discovery for each spectrum is shown in the upper-right corner of each panel, with colors matching the colors of the spectra. We have subtracted the host galaxy spectra and a locally defined low-order continuum around the lines from each spectrum. ASASSN-14li shows a He II profile that narrows over time in all epochs, while ASASSN-14ae only shows significant He II emission at later epochs. The He II profile of ASASSN-14li shows two velocity peaks, one at 0 km s−1 and one at ∼ −2000 km s−1, which evolve in their relative intensity over time.

70 3000

2500 ) 1 −

2000 FWHM (km s α H

1500

1000 0 20 40 60 80 100 120 140 Days After Discovery

Fig. 3.11.— Evolution of the Hα line width. As the luminosity of the line also decreases with time (see Figure 3.8), the line is becoming narrower as the luminosity decreases. This is the opposite of what is seen in reverberation mapping studies of quasars.

71 Filter Magnitude Uncertainty

NUV 19.77 0.06 u 17.61 0.03 g 16.15 0.03 r 15.63 0.03 i 15.37 0.03 z 15.17 0.03 J 14.30 0.07 H 13.48 0.07

Ks 13.20 0.09

These are 5′′.0 radius aperture magnitudes

SDSS (ugriz) and 2MASS (JHKS) and a PSF NUV magnitude from GALEX.

Table 3.1. Photometry of the Host Galaxy of ASASSN-14li.

72 UT Date MJD Telescope/Instrument Exposure (s)

2014 November 30.60 56991.60 UH-2.2m/SNIFS 2 × 400 2014December02.51 56993.51 APO-3.5m/DIS 2 × 1000 2014December09.51 57000.51 MDM-2.4m/Modspec 4 × 600 2014December10.56 57001.56 MDM-2.4m/Modspec 3 × 240 2014December12.48 57003.48 APO-3.5m/DIS 2 × 1200 2014December12.54 57003.54 MDM-2.4m/Modspec 1 × 600 2014 December 14.49 57005.49 FLWO-1.5m/Fast 3 × 900 2014 December 15.51 57006.51 FLWO-1.5m/Fast 3 × 1200 2014 December 19.49 57010.49 FLWO-1.5m/Fast 1 × 1800 2014December23.51 57014.51 MDM-2.4m/OSMOS 3 × 1200 2015 January 03.35 57025.35 Magellan-6.5m/IMACS 2 × 500 2015January20.53 57042.53 LBT-8.4m/MODS 2 × 900 2015February11.40 57064.40 MDM-2.4m/OSMOS 3 × 1200 2015 February 14.50 57067.50 FLWO-1.5m/Fast 2 × 1200 2015February15.38 57069.38 LBT-8.4m/MODS 3 × 1200 2015March11.35 57092.35 MDM-2.4m/OSMOS 3 × 1200 2015 April 17.19 57129.19 LBT-8.4m/MODS 3 × 900

Table 3.2. Spectroscopic Observations of ASASSN-14li.

73 MJD Hα (6563 A)˚ Hβ (4861 A)˚ He II (4686 A)˚

56993.51 1.01 × 1041 3.85 × 1040 5.07 × 1040 57001.56 1.01 × 1041 2.98 × 1040 5.53 × 1040 57003.48 1.12 × 1041 3.65 × 1040 7.09 × 1040 57003.54 8.27 × 1040 2.80 × 1040 5.45 × 1040 57005.49 9.72 × 1040 3.65 × 1040 6.06 × 1040 57006.51 7.69 × 1040 5.63 × 1040 8.67 × 1040 57010.49 9.86 × 1040 2.82 × 1040 6.52 × 1040 57014.51 6.72 × 1040 2.64 × 1040 3.12 × 1040 57025.35 6.58 × 1040 1.99 × 1040 3.77 × 1040 57042.53 4.90 × 1040 1.91 × 1040 5.07 × 1040 57064.40 2.10 × 1040 1.29 × 1040 3.79 × 1040 57067.49 2.55 × 1040 1.42 × 1040 2.38 × 1040 57069.38 4.07 × 1040 1.16 × 1040 2.91 × 1040 57092.35 1.41 × 1040 — 2.05 × 1040 57129.19 1.59 × 1040 6.96 × 1039 1.70 × 1040

Column headers indicate the lines used to measure the luminosities. All luminosities are given in units of ergs s−1. No value is given for epochs where a measurement was not possible.

Table 3.3. Line Luminosities for ASASSN-14li.

74 Vpeak t1 (days) h∆ti (days) Ng∆t (years)

−18 10 4.88 4661 20 6.31 6028 30 7.11 6793 40 7.70 7351 −19 10 20.51 19588 20 25.48 24331 30 28.53 27251 40 30.68 29307 −20 10 55.78 53280 20 67.77 64730 30 75.13 71753 40 80.60 76980 −21 10 131.82 125903 20 159.56 152400 30 176.58 168655 40 189.30 180803

For a given transient peak Vpeak and time to decay

one magnitude t1, ∆t is the average number of days per galaxy in which the transient would be detected, leading

to a total survey time of Ng∆t where Ng = 348853 is the total number of 0.01

Table 3.4. Detection Statistics for TDE Rate Calculations.

75 Chapter 4: ASASSN-15oi: A Rapidly Evolving, Luminous Tidal Disruption Event at 216 Mpc

In this chapter I describe the discovery and follow-up observations of ASASSN-15oi, the third TDE discovered by ASAS-SN.

Our transient detection pipeline was triggered on 2015 August 14 by a new source at RA/Dec = 20:39:09.12/−30:45:20.84 (J2000) with V = 16.2 ± 0.1 mag (Brimacombe et al. 2015). The source was not detected (V ∼> 17.2 mag) on 2015 July 26 or earlier. The object’s position corresponds to 2MASX J20390918-3045201, a galaxy with no previous redshift measurement. Follow-up images obtained on 2015 August 25 with the Swift UVOT roming05 and XRT (Burrows et al. 2005) conﬁrmed the detection of the transient in both the UV and X-rays.

A follow-up spectrum obtained by the Public ESO Spectroscopic Survey for Transient Objects (PESSTO; Smartt et al. 2015) using the ESO New Technology Telescope at La Silla showed that the transient exhibited a strong blue continuum and a broad (∼ 10000 km s−1) He II 4686A˚ emission feature (Prentice et al. 2015), consistent with helium-rich optical TDEs (e.g., Gezari et al. 2012; Arcavi et al. 2014). Although the redshift was originally reported as z =0.02, our later spectra showed that the redshift of the host galaxy is z = 0.0484, corresponding to a luminosity −1 −1 distance of 216 Mpc (H0 = 69.6 km s Mpc , ΩM =0.29, ΩΛ =0.71). Because of the similarities between ASASSN-15oi and previously discovered TDEs, we began a long-term follow-up campaign to monitor and characterize the object.

This chapter is adapted from “ASASSN-15oi: A Rapidly Evolving, Luminous Tidal Disruption

Event at 216 Mpc”, T. W.-S. Holoien, et al. MNRAS, 463, 3813, (2016).

76 In §4.1 I describe pre-outburst data of the host galaxy as well as new observations obtained of the transient during our follow-up campaign. In §4.2 wI analyze these data to model the transient’s luminosity and temperature evolution and compare the properties of ASASSN-15oi to those of SNe and other TDEs to determine the nature of the transient. Finally, in §4.3 we discuss our ﬁndings and how they relate to the active ﬁeld of TDE research.

4.1. Observations and Survey Data

Here I summarize the available archival survey data of the transient host galaxy 2MASX J20390918-3045201 as well as our new photometric and spectroscopic observations of ASASSN-15oi.

4.1.1. Archival Data

Because the source is located in the Southern hemisphere, there are no available archival imaging or spectroscopic data in SDSS. We obtained archival near-IR JHKs images from 2MASS (Skrutskie et al. 2006) and measured 5′′.0 aperture magnitudes for the galaxy in these images. This aperture radius, chosen to match the PSF of the Swift data, was also used to measure the source ﬂux in follow-up data. We also obtained a V -band aperture magnitude for the host galaxy by stacking several epochs of ASAS-SN data and measuring the magnitude in the same way using a 2-pixel aperture, roughly corresponding to a 14′′.0 aperture due to the large ASAS-SN pixel scale. These measured magnitudes were later used to model the host galaxy SED and subtract the host galaxy ﬂux from follow-up data of the transient. We present the measured 5′′.0 aperture magnitudes from the 2MASS and ASAS-SN images in Table 4.1.

2MASX J20390918-3045201 is not detected in archival Spitzer, Herschel, HST, Chandra, XMM-Newton, NRAO VLA Sky Survey (NVSS), Sydney University Molonglo Sky Survey (SUMSS), or VLA FIRST data. The host is also not detected in Galaxy Evolution Explorer (GALEX) UV data, but we obtain a 3-sigma upper

77 limit on the magnitude of NUV > 20.57 using co-added data totaling roughly 200s of exposure time. We use this limit when ﬁtting the host galaxy SED in §4.1.2.

Examining data from RASS (Voges et al. 1999), we do not detect the host galaxy to a 3-sigma upper limit of 2.0 × 10−2 counts s−1 in the 0.3 − 10.0 keV band. −13 −1 −2 41 −1 This corresponds to a limit of 1.2 × 10 ergs s cm (LX < 6.7 × 10 ergs s ), which provides evidence that the galaxy does not host a strong AGN. Detections of the host in WISE (Wright et al. 2010) data corroborate this picture, as the galaxy has a mid-IR (MIR) color of (W 1 − W 2) ≃ 0.06 ± 0.06, implying that any AGN activity in the galaxy is not strong (e.g., Assef et al. 2013).

Using Fast (Kriek et al. 2009), we ﬁt stellar population synthesis (SPS) models to the archival GALEX NUV limit, ASAS-SN V -band measurement, and

2MASS JHKs magnitudes of the host galaxy. The ﬁt was done assuming a Cardelli,

Clayton & Mathis (1988) extinction law with RV =3.1 and a Galactic extinction of

AV = 0.19 mag based on Schlafly & Finkbeiner (2011), an exponentially declining star-formation history, a Salpeter IMF, and the Bruzual & Charlot (2003) models. 2 We obtained a good SPS fit (reduced χν = 1.39), with the following parameters: +0.4 10 +1.8 M∗ = (1.1−0.3) × 10 M⊙, age = 2.8−1.8 Gyr, and a 1σ upper limit on the star −1 formation rate of SFR ≤ 0.002 M⊙ yr . Scaling the stellar mass of 2MASX J20390918-3045201 using the average stellar-mass-to-bulge-mass ratio from the hosts of ASASSN-14ae and ASASSN-14li (Holoien et al. 2014, 2016a) implies a bulge mass 9.8 of MB ≃ 10 M⊙. Using the MB − MBH relation from McConnell & Ma (2013), we 7.1 obtain a black hole mass of MBH = 10 M⊙, which is significantly more massive than those of the hosts of the previous ASAS-SN TDEs. Fits to the transient spectral energy distribution give no indication of additional extinction related to the host galaxy. In the analyses of the event’s SED which follow, we only correct for Galactic extinction.

As there were no archival data in many of our follow-up filters, we lack pre-outburst magnitudes of the host. In order to obtain host magnitudes to use for host flux subtraction, we produced synthesized Swift UVOT and Bessel BVI host magnitudes using the best-fit Fast SED. We used a bootstrapping scheme to assess the uncertainties associated with our determination of the host magnitudes.

78 We generated 1000 mock input SEDs based on the observed ﬂuxes and assuming Gaussian errors, and constructed distributions of the magnitudes synthesized from this ensemble. Our adopted host magnitudes fall within ∼< 0.2 mag of the peak of the bootstrapped distributions. These estimated host magnitudes are listed in Table 4.2.

4.1.2. New Photometric Observations

After the transient was classified as a TDE, we obtained a series of 26 Swift XRT and UVOT target-of-opportunity (ToO) observations. The UVOT observations were obtained in 6 filters: V (5468 A),˚ B (4392 A),˚ U (3465 A),˚ UV W 1 (2600 A),˚ UV M2 (2246 A),˚ and UV W 2 (1928 A)˚ (Poole et al. 2008). We extracted source counts from a 5′′.0 radius region and sky counts from a ∼40′′.0 radius region using the UVOT software task Uvotsource and converted these count rates into magnitudes and fluxes using the most recent UVOT calibrations (Poole et al. 2008; Breeveld et al. 2010).

The XRT operated in Photon Counting mode (Hill et al. 2004) for our observations. The data from all epochs were reduced with the software task Xrtpipeline. We extracted X-ray source counts and background counts with Xselect using a region with a radius of 10 pixels (23′′.6) centered on the source position and a source-free region with a radius of 100 pixels (235′′.7), respectively. While the source is undetected in most individual epochs, we detect X-ray emission consistent with the position of the transient after combining the signal from multiple exposures. We assume a blackbody plus power law model (see §4.2.1) and Galactic H I column density of 5.6×1020 cm−2 (Kalberla et al. 2005) to convert the detected counts into ﬂuxes, with a count to ﬂux conversion factor of 5.2 × 10−11 ergs s−1 (cm2 counts s−1)−1 for the absorbed spectrum and 8.5 × 10−11 ergs s−1 (cm2 counts s−1)−1 for the absorption-corrected spectrum.

We also obtained BVI images with the LCO (Brown et al. 2013) 1-m telescopes at Siding Spring, South African Astronomical, and Cerro Tololo Inter-America Observatories. We measured magnitudes using aperture photometry with a 5′′.0

79 aperture radius in order to match the host galaxy and Swift measurements. Photometric zero-points were determined using several stars in the ﬁeld.

In order to constrain any offset between the source of the outburst and the nucleus of the host galaxy, we astrometrically aligned an I-band image of the transient taken on 2015 September 4 with the LCOGT 1-m telescope at Sutherland, South Africa with the archival DSS image of the host galaxy. We then measured the offset between the centroid position of the transient in the LCOGT image and that of the host galaxy in the DSS image, finding an offset of 0.24 ± 0.05 arcseconds (254.8 ± 53.1 parsecs). We performed the same analysis using two LCOGT V -band images from 2015 September 4 and 2015 November 23, after the transient had faded. In order to increase the accuracy of the transient position, we first used Hotpants1, an implementation of the Alard (2000) algorithm for image subtraction, to subtract the 2015 November 23 template image from the 2015 September 4 image to isolate the transient flux. The centroid position of the transient in the subtracted image is located 0′′.08 West and 0′′.05 North of the centroid position of the host in the template image, giving a total offset of 0.09 ± 0.06 arcseconds (96.3 ± 62.8 parsecs). Figure 4.1 shows the two LCOGT V -band images used to measure the offset. Higher resolution data, particularly in the UV, could improve these estimates.

Figure 4.2 shows the X-ray, UV, and optical light curves of ASASSN-15oi. Table A.4 and Table A.5 give the XRT flux measurements and UVOT/V RI magnitudes, respectively. The photometric observations cover the period from MJD 57248.2 (the epoch of discovery) through our latest epoch of observations on MJD 57355.0, a span of 106.8 days. The data shown in Figure 4.2 are presented without extinction correction or host flux subtraction, and the Swift B- and V -band data were converted to the Johnson-Cousins system using the color corrections found at http://heasarc.gsfc.nasa.gov/docs/heasarc/caldb/swift/docs/uvot/. Also shown in Figure 4.2 are the host magnitudes estimated from the host SED fit. Our observations show that ASASSN-15oi brightened considerably with respect to the host galaxy in the UV and bluer optical filters in the short period between our previous observation on 2015 July 25 and our detection of the transient on 2015

1http://www.astro.washington.edu/users/becker/v2.0/hotpants.html

80 August 14. The largest increase came in the Swift UV M2 band, where it brightened by ∆mUVW 2 ∼ −6.8, while the V -band increase was weaker, at ∆mV ∼ −1.2. This large UV increase is additional evidence against typical AGN variability, which is normally of much smaller magnitude (e.g., Gezari et al. 2012; MacLeod, Guillochon, & Ramirez-Ruiz 2012).

4.1.3. New Spectroscopic Observations

We obtained spectra of ASASSN-15oi spanning 69 days between UT 2015 September 04 and UT 2015 November 07. The spectrographs used for these observations were the Ohio State Multi-Object Spectrograph (OSMOS; Martini et al. 2011) mounted on the MDM Observatory Hiltner 2.4-m telescope (4200 − 6800 A,˚ R ∼ 4 A),˚ the Fast Spectrograph (FAST; Fabricant et al. 1998) mounted on the Fred L. Whipple Observatory Tillinghast 1.5-m telescope (3700 − 9000 A,˚ R ∼ 3 A),˚ and the Wide Field Reimaging CCD Camera (WFCCD) mounted on the Las Campanas Observatory du Pont 2.5-m telescope (3700 − 9500 A,˚ R ∼ 7 A).˚ The spectra were reduced using standard techniques in IRAF, and we applied telluric corrections using observations of spectrophotometric standard stars from the same nights. Each spectrum was scaled to match the V -band photometry. Figure 4.3 shows a time-sequence of the ﬂux-calibrated follow-up spectra as well as the initial spectrum obtained by PESSTO (Prentice et al. 2015). Summary information for each spectrum is listed in Table 4.3.

The key characteristics of the early spectra of ASASSN-15oi are a strong blue continuum, consistent with the photometric measurements, and the presence of broad helium lines in emission. The emission features are broad and asymmetric, with initial widths of ∼ 10000 − 20000 km s−1 that narrow over time. In the initial spectrum obtained by PESSTO, the helium lines are blueshifted by approximately 6000 − 8000 km s−1, but in later spectra little-to-no apparent shift relative to the systemic velocity is observed. The transient features fade rapidly, with spectra later than 2015 September 12 showing no blue continuum and absorption features consistent with the transient’s host galaxy. Unlike other recently discovered TDEs (e.g. Arcavi et al. 2014; Holoien et al. 2016a; French, Arcavi, & Zabludoﬀ 2016),

81 the later host-dominated spectra do not show features, such as Hδ absorption with a large equivalent width, that are consistent with the host being a post-starburst galaxy. We further analyze the features of these spectra and compare them to other TDEs and SNe in §4.2.

4.2. Analysis

4.2.1. SED Analysis

Combining all 61.4 ks of observations, we were able to extract an X-ray spectrum of the source. The spectrum was analyzed with Xspec 12.8.2 (Arnaud 1996) using Cash statistics (Cash 1979). A fit with a single power law model does not result in an acceptable fit. The X-ray spectral slope is very steep (Γ=5.8) with large residuals at energies above 2.0 keV. We found an acceptable fit using an absorbed blackbody plus power law model with the absorption column density fixed to the +10 Galactic value. This results in a blackbody temperature at an energy of 49−9 eV +1.5 and a photon index for the hard energy component of Γ = 1.8−0.8. The absorbed +0.4 −14 −1 −2 and unabsorbed fluxes in the 0.3-10 keV band are 8.0−0.3 × 10 ergs s cm and +0.9 −14 −1 −2 18.7−0.6 × 10 ergs s cm , respectively. The X-ray spectrum and the model fits are shown in Figure 4.4.

There are two challenges to interpreting the X-ray detections. First, the archival ROSAT data are not deep enough to rule out a pre-existing source with the presently 41 41 observed X-ray luminosity (LX < 6.7 × 10 versus LX =4.8 × 10 ergs/s). Second, the X-ray emission we observe during the transient is consistent with constant luminosity. We can rule out X-ray emission dominated by sources other than an AGN. Low Mass X-ray Binaries (LMXBs) in old stellar populations, like the Fast models for the SED of 2MASX J20390918-3045201, produce X-rays. Based on the correlations of Kim & Fabbiano (2004) between the K-band luminosities of galaxies and the integrated X-ray emission by their LMXBs, however, this contribution should only be of order 1040 ergs s−1, well below the observed luminosity. Thus, the X-ray emission is almost certainly dominated by AGN activity, and seems likely to

82 be associated with the present transient, but this association is not required by the available data. Late-time observations of the X-ray evolution should illuminate the source of this emission.

Figure 4.5 shows the evolution of the UV-optical SED of ASASSN-15oi. Using the 5′′.0 aperture host magnitudes estimated from the host SED fit, we produced host- and extinction-corrected light curves for all optical and UV filters. We then fit these host-subtracted fluxes with blackbody models using MCMC methods, as was done for the previous ASAS-SN TDEs (Holoien et al. 2014, 2016a). Due to there being uncertainty in the actual host magnitude, we only included epochs for which the transient’s flux was greater than 50% of the host galaxy flux when performing these SED fits. This criteria only excludes later epochs of observations in optical filters, as the UV emission remains well above our estimated host levels throughout our observations.

Unlike the previous ASAS-SN TDEs, the data for ASASSN-15oi indicate that the source’s temperature increases before (probably) leveling off in later epochs. Because of this, we fit the temperature using a changing prior based on initial, unconstrained fits. For epochs within 10 days of discovery, the temperature was fit with a prior of T = 2 × 104 K. For later epochs, we fit the data with a prior of T = (2+(∆t − 10)/2) × 104 K, where ∆t = MJD − 57248.2 and the temperature is capped at T = 40000 K, as this provides a better fit than allowing the temperature to increase arbitrarily and is consistent with the color evolution (see §4.2.3). These priors were applied with a log uncertainty of ±0.05 dex in all epochs. Due to the fact that our Swift data do not span the peak of the SED for the later, hotter epochs, the temperatures in these phases are not well-constrained, although the prior used for these fits does capture the steepening spectral slope. Also shown in Figure 4.5 are the X-ray spectral model (see Figure 4.4) and the ionizing luminosity implied by the He II 4686A˚ line luminosity (see §4.2.2). It can be clearly seen that the blackbody model inferred from the optical/UV data is not enough to explain the observed X-ray and line emission.

The optical/UV luminosity of ASASSN-15oi, shown in Figure 4.6, fades steadily over the ∼ 3 months after the initial discovery with the temperature constrained as

83 described above. As was the case with previous ASAS-SN TDEs, the luminosity evolution inferred from the blackbody fit at early times can be well-described by an −(t−t0)/τ exponential L = L0e (Holoien et al. 2014, 2016a), with best-fit parameters of log L0/L⊙ = 10.8 and τ ≃ 46.5 days for t0 = 57241.6. This decay rate is quite similar to that observed for ASASSN-14ae (τ ≃ 39 days for t0 = 56682.5). However, the late-time light curves of ASASSN-14ae and ASASSN-14li can also be fit by power-law profiles (Brown et al. 2016; Brown et al. 2017). For this reason, we also −5/3 fit the luminosity evolution of ASASSN-15oi with the popular L = L0(t − t0) power law, finding a best fit of log L0/L⊙ = 13.3 and t0 = 57228.0, although this is a significantly worse fit than the exponential (χ2 = 55.4 vs. χ2 = 15.1). For −5/3 the t power law model, we used the constraint 57228 ≤ t0 ≤ 57248 so that the disruption begins sometime between our previous non-detection and the discovery of the transient. If we relax this constraint and allow t0 and the power law index to vary freely, we find a best fit of log L0/L⊙ = 13.3 and t0 = 57212.3 with a power law index of α = 1.62. This model is still a poorer fit than the exponential model (χ2 = 19.0 vs. χ2 = 15.1), and it would require a fairly rapid rise to be consistent with our flux limits from numerous epochs of data obtained between MJD 57212 and MJD 57228. With a longer temporal baseline, we will be better able to distinguish between exponential and power-law decline models (e.g., Brown et al. 2016).

This temperature and luminosity evolution is inconsistent with what would be expected if ASASSN-15oi were an SN, which typically exhibit rapidly declining temperatures along with constant or declining luminosity (e.g., Miller et al. 2009; Botticella et al. 2010; Inserra et al. 2013; Graham et al. 2014). Normal AGN activity is ruled out by the archival host photometry, leaving a TDE as the most likely explanation for ASASSN-15oi.

In contrast to the optical/UV luminosity evolution, the X-ray luminosity remains constant for the duration of the observations. However, even in later epochs, the X-ray luminosity is nearly two orders of magnitude lower than the optical/UV luminosity, and even if it is associated with the TDE, it has little impact on the overall energy output of the source. As shown in Figure 4.5, a single blackbody model is incapable of ﬁtting both the X-ray and optical/UV emission. Thus, if the

84 X-ray emission is associated with the TDE, we infer that the X-ray emission likely comes from a diﬀerent region of the TDE than the optical/UV emission. Including the UV luminosity implied by the He II emission, ASASSN-15oi is another piece of evidence pointing to the need for more complex models of TDE emission (e.g., Metzger & Stone 2016; Roth et al. 2016).

Integrating the combined optical/UV and X-ray luminosity curves gives a total radiated energy of ∼ 6.6 × 1050 ergs for the ∼ 3.5 months of follow-up observations. The uncertainties in this estimate are predominantly systematic due to the poorly constrained temperatures at later times.

4.2.2. Spectroscopic Analysis

Though the SED evolution of ASASSN-15oi is largely consistent with the transient being a tidal disruption ﬂare, the rapid evolution and absolute magnitude are unique among TDEs discovered at optical wavelengths, and it is worth examining other possible sources of the observed emission. The most likely alternative explanation for the transient is that it was an SN, as archival X-ray, UV, infrared, and radio observations do not show any indication that 2MASX J20390918-3045201 hosts a typical AGN. While SNe evolve on the rapid timescales observed for ASASSN-15oi, a spectroscopic comparison between ASASSN-15oi and SNe reveals that this is not the case.

ASASSN-15oi shows broad He emission lines in its early spectra but little-to-no H emission. If ASASSN-15oi were an SN, it would likely be a Type Ib or Type Ic-bl: the former because SNe Ib show He emission with no H emission, and the latter because broad-line SNe Ic exhibit line widths comparable to those observed in ASASSN-15oi. Figure 4.7 shows a comparison between spectra taken at or before maximum for ASASSN-15oi (Prentice et al. 2015), the He-rich TDE PTF09ge (Arcavi et al. 2014), the Type Ib SN SN 2004gq (Modjaz et al. 2014), and the broad-line Type Ic SN SN 2002ap (Gal-Yam, Ofek & Shemmer 2002). While SN 2004gq does show the helium emission lines present in the spectrum of ASASSN-15oi, there are also numerous absorption features that are not present in the spectrum of

85 ASASSN-15oi. Similarly, SN 2002ap shows broad emission features, but does not have the same helium emission lines, and the shape of the continuum is very diﬀerent. In contrast, the spectrum of PTF09ge is very similar to that of ASASSN-15oi, as both objects show a strong blue continuum, a broad He II 4686A˚ emission feature, and no hydrogen emission. Given this comparison, we conclude that it is very unlikely that ASASSN-15oi is a SN.

Arcavi et al. (2014) noted that optically-selected TDEs span a continuum from H-rich to He-rich spectroscopic features. As indicated in Figure 4.7, early follow-up spectra of ASASSN-15oi are highly consistent with those of He-rich TDEs like PS1-10jh and PTF09ge (Gezari et al. 2012; Arcavi et al. 2014). The prominent features of the earliest spectra are a strong blue continuum with broad He II and He I emission features. The earliest spectrum, obtained by PESSTO on 2015 August 21, shows He II 4686A˚ and He I 5876A˚ emission −1 −1 features with FWHMHeII ≃ 19700 km s and FWHMHeI ≃ 18800 km s , respectively, that are blueshifted by ∼ 6000 − 8000 km s−1. The second spectrum, obtained with the du Pont+WFCCD on 2015 September 4, shows some interesting changes. The He II 4686A˚ feature is still prominent but has a smaller width −1 (FWHMHeII ≃ 9200 km s ) compared to the first spectrum and the peak of the line is only blueshifted by ∼ 1400 km s−1. Conversely, the He I 5876A˚ line has −1 remained broad, with FWHMHeI ≃ 18800 km s . There are two new broad bumps (FWHM ∼ 10000 km s−1) that appear around the strong He II 4686A˚ line with peak wavelengths at ∼ 4435A˚ and ∼ 5000A.˚ These features are consistent with the He I 4472A˚ and He I 5016A˚ transitions, blueshifted by 1000 − 2500 km s−1. This spectrum also shows a new feature at ∼ 6600A˚ which exhibits an integrated flux that is comparable to the He II 4686A˚ line flux, indicated with a red dotted line in Figure 4.3. This is near the position of Hα, but it is likely that this feature is due at least in part to the He II 6560A˚ n =6 → 4 line, which is expected to be an order of magnitude weaker than He II 4686A˚ in photoionized gas (Osterbrock 1989; Gezari et al. 2015).

Unlike what has been seen in previous TDEs, the spectroscopic features of ASASSN-15oi evolve rapidly. Between the classiﬁcation spectrum obtained on UT

86 2015 August 21 and the ﬁrst follow-up spectrum obtained on UT 2015 September 4, the blue continuum in the spectra of ASASSN-15oi fades and the He II line narrows considerably. The emission lines seem to have completely faded by the time of the latest follow-up spectrum obtained on UT 2015 November 7, slightly less than 3 months after discovery.

After correcting for Galactic reddening and subtracting an estimated continuum from the PESSTO classification spectrum, we estimate the luminosity of the He II 42 −1 4686A˚ line to be LHeII ≃ (1.1 ± 0.1) × 10 ergs s , where the dominant source of uncertainty is in setting the continuum level. In the UT 2015 September 4 spectrum, 41 −1 the line has faded to a luminosity of LHeII ≃ (1.7 ± 0.5) × 10 ergs s . We are unable to obtain estimates of the line luminosity in later spectra, as the line has faded and the continuum is too uncertain to provide an accurate subtraction. Assuming case B recombination (Osterbrock 1989), the line luminosity from the classification 42 −1 spectrum implies an ionizing luminosity of Li ≃ (7.3 ± 0.7) × 10 ergs s , while the line luminosity from the September 4 spectrum implies an ionizing luminosity 42 −1 of Li ≃ (1.1 ± 0.3) × 10 ergs s . While this is a rapid drop, in both cases the line luminosity is greater than the peak line luminosity measured for ASASSN-14li (Holoien et al. 2016a). As shown in Figure 4.5, the optical/UV emission is unable to provide the requisite ionizing luminosity needed to power the observed He II line, and additional ionizing flux must be coming from a different region of the TDE.

It is not uncommon for the emission features in TDE spectra to narrow and become less luminous over time (e.g., Holoien et al. 2014, 2016a), which is the opposite behavior to that observed in reverberation mapping studies of quasars, where the line width broadens as the luminosity decreases (e.g. Peterson et al. 2004; Denney et al. 2009). However, while the spectroscopic evolution of ASASSN-15oi is similar to that of other optically-selected TDEs, it occurs on a signiﬁcantly faster timescale. This rapid line evolution could be indicating that the “reprocessing layer” responsible for the optical/UV emission and emission lines, is becoming optically thin or is no longer in our line-of-sight to the TDE. It is also worth noting that ASASSN-14ae, which had a similar ∼ 3 week gap between the previous epoch of observation and the epoch of discovery, also showed a blueshift in its broad

87 emission features in its early spectra which was not present in later spectroscopic observations (Holoien et al. 2014). While ASASSN-14li did not show any blueshift in its spectroscopic observations (Holoien et al. 2016a), it also had a signiﬁcantly longer gap (∼ 3 months) between the epoch of discovery and the previous epoch of observations due to Sun constraints. It is possible that an early blueshift in the emission lines of optically-selected TDEs is common, and that it was not seen in ASASSN-14li due to the gap in observations.

As discussed by French, Arcavi, & Zabludoff (2016), optically-selected TDEs have shown a preference for unusual “quiescent Balmer-strong galaxies”—those galaxies whose spectra show little-to-no line emisison but strong Balmer line absorption, indicating a recent period of intense star formation. While archival spectra of 2MASX J20390918-3045201 are not available, our later epochs of follow-up spectra are dominated by the host galaxy, allowing us to determine whether the host of ASASSN-15oi also falls into this rare class of galaxies. From the 2015 November 7 du Pont/WFCCD spectrum, we measure Hδ and Hα absorption features with EW= 1.1 A˚ and EW=0.6 A,˚ respectively. In French, Arcavi, & Zabludoff (2016), a quiescent Balmer-strong galaxy is defined as having absorption with a Lick HδA index of HδA > 4 A˚ (equivalent to Hδ EW > 3 A)˚ and Hα EW < 3 A.˚ With much narrower Hδ absorption, and Hα in absorption rather than emission, 2MASX J20390918-3045201 clearly does not fall into the same class of galaxy as the TDE hosts discussed in French, Arcavi, & Zabludoff (2016), and is more consistent with the old stellar population indicated by the Fast SED fits.

4.2.3. Comparison of ASAS-SN TDEs

ASAS-SN has now found three nearby TDEs, all of which were intensively observed with Swift and ground-based telescopes for months after discovery, allowing us to compare their photometric properties and the inferred temperatures and luminosities at early and late epochs. Such an analysis was not possible with many earlier TDEs, due to the fact that they were fainter, and late-time observation in many cases was not possible. For optical TDEs where such analyses were possible, previous studies have shown that they typically show little-to-no color evolution or change

88 in temperature (e.g., van Velzen et al. 2011; Gezari et al. 2012; Arcavi et al. 2014; Chornock et al. 2014), and while we find that the first two ASAS-SN TDEs follow this pattern, ASASSN-15oi appears to be different. All data for ASASSN-14ae and ASASSN-14li shown below are taken from Holoien et al. (2014) and Holoien et al. (2016a), respectively.

From the host-subtracted light curves we produced color evolution curves for ASASSN-15oi, ASASSN-14ae, and ASASSN-14li. Figure 4.8 compares the Swift (UV M2 − UV W 2) and (UV M2 − U) colors of the three TDEs. All three objects show blue UV-UV and UV-optical colors in all epochs, which is expected for optically-selected TDEs. The (UV M2 − UV W 2) colors of all three objects are remarkably similar and show little evolution for all three objects, but the (UV M2 − U) colors do show some variation: while they remain blue overall, ASASSN-14ae becomes slightly redder and ASASSN-15oi becomes slightly bluer before leveling oﬀ in later epochs, while ASASSN-14li shows little evolution. The blue evolution of the (UV M2 − U) color for ASASSN-15oi is in line with the increasing temperature described in §4.2.1, and it is not unexpected that ASASSN-15oi would diﬀer slightly from the other two TDEs, as ASASSN-14ae and ASASSN-14li both had roughly constant temperatures in all epochs. In all cases, the magnitude of the color change is fairly minor compared to that of SNe, which typically become redder much more rapidly (see Figure 7 of Holoien et al. (2014) for a comparison).

Figure 4.9 shows a comparison of the inferred blackbody SEDs for all three objects from epochs shortly after discovery. In all three cases the ﬁt shown used both Swift and ground-based data, and for ASASSN-14li and ASASSN-15oi the data were ﬁt assuming a temperature prior. ASASSN-14ae and ASASSN-15oi exhibited similar early temperatures of roughly T ∼ 20000 K while ASASSN-14li was hotter, with T ∼ 35000 K. Though ASASSN-15oi is more luminous and evolved more rapidly than the other two ASAS-SN TDEs, its inferred blackbody temperature is not unusual for an optically-selected TDE.

Finally, we compare the blackbody luminosity, temperature, radius, and emission line evolution inferred for the three ASAS-SN TDEs in Figures 4.10, 4.11, 4.12, and 4.13, respectively. In all cases the bolometric luminosity was inferred from

89 the optical/UV SED ﬁts, as described in §4.2.1, and does not include the X-ray emission detected for ASASSN-15oi or ASASSN-14li. ASASSN-15oi is considerably more luminous than the other two TDEs, but its bolometric luminosity falls at a fairly similar rate to that of ASASSN-14ae. ASASSN-14li fades much more slowly than the other two; while it initially has a similar luminosity as ASASSN-14ae, by later epochs it is nearly an order of magnitude more luminous than ASASSN-14ae, and ASASSN-15oi has fallen to a comparable luminosity. This is perhaps correlated with the mass of the black hole involved in the TDE: though the three ASAS-SN TDEs have similar black hole masses given the uncertainties on the mass estimates, 6.7 ASASSN-14li ostensibly had the least massive black hole, at MBH ∼ 10 M⊙, while 6.8 ASASSN-14ae and ASASSN-15oi had black holes of masses MBH ∼ 10 M⊙ and 7.1 MBH ∼ 10 M⊙, respectively.

Including both apparently thermal optical/UV and X-ray emission (where observed), we estimate total radiated energies of 1.7 × 1050 ergs, 7.0 × 1050 ergs, and 6.6 × 1050 ergs for ASASSN-14ae, 14li, and 15oi, respectively. Converted to an equivalent rest mass, these energies correspond to 0.001, 0.004, and 0.004 M⊙, compared to a typical TDE stellar mass of M∗ = 0.3M0.3M⊙ (Kochanek 2016b). 2 If the radiated energy is E = ηfM∗c where η ≃ 0.1 is the radiative eﬃciency of accretion and f is the fraction of the stellar mass accreted (f ≃ 0.5 is bound to the −1 −1 −1 black hole), then we must have ηf ≃ 0.003M0.3 , 0.013M0.3 and 0.012M0.3 for these three events. All three events require either that TDE activity is less radiatively eﬃcient than normal disk accretion (η ≪ 0.1) or that a very small fraction of the bound mass (f ≪ 0.5) is accreted. Variants of these possibilities appear in recent theoretical models (e.g., Metzger & Stone 2016; Piran et al. 2015; Strubbe & Murray 2015; Svirski, Piran & Krolik 2017).

ASASSN-15oi stands out very clearly from the other two ASAS-SN TDEs in its temperature evolution. While ASASSN-14ae and ASASSN-14li evolve at nearly constant temperatures, ASASSN-15oi increases in temperature. The exact increase is poorly constrained because the data do not sample the peak of the SED, but it is consistent with a model where the temperature linearly increases from T ∼ 20000 K to T ∼ 40000 K over roughly 50 days and then remains roughly constant thereafter.

90 ASASSN-15oi was fit with the temperature prior described in §4.2.1 above and ASASSN-14li was fit with a prior of log T/K =4.55 ± 0.05, as described in Holoien et al. (2016a). As the UV data for ASASSN-14ae captured the peak of the transient’s SED, its temperature was not fit with a prior for those epochs where Swift data were available, and was fit with a prior that tracks the UV data for those epochs where Swift data was not available. We emphasize that there are systematic uncertainties in these temperature fits, particularly for ASASSN-15oi and ASASSN-14li, where the fits were made with very strong priors due to the lack of FUV data, resulting in the lower scatter and larger uncertainties seen in Figure 4.11. However, these fits are consistent with the data, and the early color evolution of ASASSN-15oi makes it clear that it has an increasing temperature in early epochs, where the other two ASAS-SN TDEs do not.

Figure 4.12 shows the evolution of the apparent photospheric radii of the ASAS-SN TDEs. All three show the same decreasing trend, but the rate at which the radius falls differs significantly between the three TDEs, with ASASSN-15oi standing out from the other two events. To put the radius estimates on a more physical scale, Figure 4.12 shows the radii in units of the gravitational radius 2 12 7 rg = GMBH /c = 1.5 × 10 cm of a MBH = 10 M⊙ black hole. The mass 6.8 6.7 7.1 estimates of MBH ∼ 10 , 10 and 10 M⊙ for ASASSN-14ae, 14li and 15oi are sufficiently close to this mass to use a common scaling given the overall dynamic range. For comparison, we also show the tidal disruption radius of the Sun, 1/3 13 RT = R⊙(MBH /M⊙) ≃ 1.5 × 10 cm ≃ 10rg, and the radius corresponding 1/2 to expansion at v0(r) = c(2rg/r) and initiated at the origin at t = 0. To facilitate comparisons to velocities, we also convert the radial scale on the left to the corresponding velocity scale v0 on the right.

Broadly speaking, the initial photospheric radii are close to the radius that would be achieved by free expansion at v0(r) from the disruption radius RT rather than being comparable to RT or smaller. Given the uncertainties in the black hole masses and the temporal start of the transients, this is mainly a qualitative observation. The apparent photospheres then shrink, but are still well outside RT at the end of the phase where they can be well-estimated. That the photospheric

91 radii are all large compared to RT favors models where the optical/UV emission is reprocessed by debris on large scales rather than direct emission from a disk. Piran et al. (2015) and Svirski, Piran & Krolik (2017) propose generating the luminosity by shocks at apocenter, rather than by reprocessing disk emission, but this would imply a much smaller emitting area than 4πr2 at the apocentric radius. If the covering fraction of the shock emission region is f 2, then the radius has to increase to r/f to produce the observed luminosity and this would initially lie well outside the radius corresponding to an orbit expanding at v0(r) if f ∼ 0.1, contrary to what 7 is expected for the self-intersection radius of tidal streams associated with a 10 M⊙ black hole (Dai, McKinney & Miller 2015). If the dominant source of line emission was closely associated with the photosphere, we would expect the typical line widths to steadily increase with time. As we noted earlier, this is also the pattern observed in AGN, where line widths increase with decreasing luminosity.

For comparison to the photospheric evolution, Figure 4.13 shows the evolution of the FWHM of the most prominent emission line in each TDE (Hα for ASASSN-14ae and ASASSN-14li, and He II 4686A˚ for ASASSN-15oi). The scales and limits of Figure 4.13 are the same as for Figure 4.12 except for relabeling the circular 1/2 velocity to be the FWHM (i.e. just taking v0 = c(2rg/r) = FWHM) to allow qualitative comparisons of the radial and velocity scales. For ASASSN-14ae and ASASSN-14li, the initial line formation region is of order the photospheric radius. For ASASSN-14li, the two scales are quite diﬀerent. The simplest explanation for the mild conﬂict with expansion at velocity v0(r) would be to shift the start of the expansion to be 10-20 days earlier. In all three cases, the FWHM diminishes with time, moving outwards in radius if we can interpret the velocities kinematically. At late times, the dominant source of line emission appears to arise from radii far outside the (continuum) photosphere. Though we are only able to obtain two clear measurements of the He II 4686A˚ line for ASASSN-15oi, the data show that the line is narrowing much more rapidly than Hα did for the previous two ASAS-SN TDEs. The primary caveat for the line evolution is that continuum subtraction is challenging for all these events, particularly at late times, and an additional (but sub-dominant) broad emission component might be hard to identify.

92 The extensive optical and UV data sets obtained for the three ASAS-SN TDEs allows us to perform comparisons between them that are largely impossible with other, fainter TDEs. While the three TDEs show unique characteristics, their colors, temperatures, and luminosity evolution are all broadly similar.

4.3. Discussion

ASASSN-15oi was discovered by ASAS-SN on 2015 August 14 and had a peak absolute V -band magnitude of MV ∼ −20.5. Follow-up data indicate that it is consistent both photometrically and spectroscopically with previously discovered He-rich, optically-selected TDEs, and is inconsistent with either an SN or an AGN event. Unlike other optically discovered TDEs, ASASSN-15oi faded rapidly in the optical and UV, and spectra obtained ∼ 3 months after discovery indicate that the transient features had almost completely disappeared. If we fit the optical and UV emission with a single blackbody model, the best-fit temperature is T ∼ 20000 K, and the temperature becomes hotter in later epochs. The early luminosity declines steadily at a rate best fit by an exponential decay, L ∝ e−(t−t0)/τ , with τ ≃ 46.5 days, −5/3 but is also marginally consistent with a power law decline, L ∝ (t − t0) . Like ASASSN-14li (Holoien et al. 2016a), ASASSN-15oi exhibits soft X-ray emission, though this emission is weaker than previous limits and cannot be conclusively tied to the TDE flare. Late-time X-ray observations to see if the X-ray emission starts to fade are needed to confirm the association of the X-ray emission with the transient. ASASSN-15oi is the third nearby TDE discovered by ASAS-SN, and possibly the second to exhibit both X-ray and optical/UV emission.

Early follow-up spectra show a strong blue continuum and broad He II 4686A˚ and He I 5725A˚ emission features, and while the spectroscopic features evolve rapidly, their evolution is consistent with that of a TDE, and not that of an SN or quasi-steady state AGN. These spectral features are characteristic of He-rich TDEs like PS1-10jh and PTF09ge (Gezari et al. 2012; Arcavi et al. 2014), making ASASSN-15oi another member of this intriguing class of objects which exhibit strong helium emission lines but no corresponding hydrogen emission. That He-dominated spectra are so common essentially rules out the possibility that they are produced by

93 the disruption of helium stars. While it might be possible to explain a single event such as PS1-10jh with this scenario (Gezari et al. 2012; Strubbe & Murray 2015), it seems improbable that He stars can represent ∼ 1/3 of the overall optical/UV TDE rate (see Kochanek 2016b).

In Holoien et al. (2016a), we found that a 90% confidence interval on the implied TDE rate given the ASAS-SN discoveries was r = (2.2 − 17.0) × 10−5 yr−1 per galaxy, a range that is consistent with theoretical estimates (e.g., Stone & Metzger 2016; Kochanek 2016b) while also significantly higher but consistent with the rates inferred from other optical surveys (e.g., van Velzen & Farrar 2014), given the uncertainties. The discovery of a third ASAS-SN TDE less than a year after the previous one implies that the relatively high (compared to previous observational studies) rate from Holoien et al. (2016a) is correct, which helps to alleviate some of the tension between observed and theoretical rates. Including this third TDE, ASAS-SN is finding roughly 1 TDE for every 70 type Ia SNe (3 and 211, respectively, at the time of writing), a rate that is significantly higher than that of other surveys (see Holoien et al. (2016a) for a detailed comparison). This suggests that the ASAS-SN TDE sample is significantly more complete than previous surveys, and consequently that our rate estimate may be more accurate.

ASAS-SN has now discovered three of the nearest optically discovered TDEs, each of which has been extensively observed over long periods of time in the optical, UV, and X-rays. Due to the nearby and bright nature of objects discovered by ASAS-SN, future TDEs will be similarly easy to observe with a wide variety of instruments, which will allow the creation of a catalog of well-studied TDEs that can be used to study the early and late-time behaviors of these transients, which cannot be done with higher-redshift objects. Moreover, due to the rapid evolution of the black hole mass function, the TDE rate should decline rapidly with increasing redshift (Kochanek 2016b), making TDEs creatures of the local universe probed by ASAS-SN. Given the success of the ASAS-SN TDE search to-date, we expect it will remain a powerful tool for ﬁnding and studying TDEs and other bright transients in the future.

94 Fig. 4.1.— V -band images of ASASSN-15oi shortly after discovery and after the transient had faded. The left panel shows the LCOGT 1-m image from 2015 September 4 and the right panel shows the LCOGT 1-m image from 2015 November 23. The red circles have radii of 3′′.0 and are centered on the position of the transient measured in the September 4 image.

95 12 UVW2−2.0 UVW1−0.6 B+0.9 I+5.1 UVM2−1.2 U+0.3 V+2.7 XRT 13 10−10 14

16 10−11 /s) 2

constant 18 −12 + 10 19 mag

20 X-ray Flux (ergs /cm

21 10−13

23 10−14 24 -20 -10 0 10 20 30 40 50 60 70 80 90 100110 Days Since Discovery (MJD=57248.2)

Fig. 4.2.— Light curves of ASASSN-15oi beginning on MJD= 57248.2 and spanning 107 days. Follow-up data obtained from Swift (X-ray, UV, and optical) are shown as circles and follow-up data obtained by the LCOGT 1-m telescopes (optical) are shown as squares. All magnitudes are in the Vega system (left scale), and X-ray fluxes are shown in ergs cm−2 s−1 (right scale). The scales are chosen so that time variability has the same meaning for both the X-ray and optical/UV data. The X-ray data points are averages over multiple epochs. The data are not corrected for extinction and error bars are shown for all points. The previous V -band non-detection from ASAS-SN is plotted as a triangle at −20 days, the date of our previous observation of the host galaxy. Estimated 5′′.0 aperture magnitudes measured from the host galaxy SED fit (stars) and an X-ray upper limit from RASS (triangle) are shown at +114 days. Vertical bars at the bottom of the figure indicate the dates of spectroscopic observations. Table A.5 contains all the follow-up photometric data.

96 3.0 PESSTO 2015/08/21 WFCCD 2015/09/04 OSMOS 2015/09/12 2.5 FAST 2015/10/11 He II FAST 2015/10/12

˚ WFCCD 2015/10/17 A) /

2 WFCCD 2015/11/07 2.0

1.5 ergs/s/cm 15

− He I 10 ( 1.0 6600A˚ λ f

0.5

0.0 4000 5000 6000 7000 8000 Rest Wavelength (A)˚

Fig. 4.3.— Spectroscopic time-sequence of ASASSN-15oi. The dates of the observations and the instruments used are listed in matching colors and order to the spectra. (See Table 4.3 for more information.) The prominent He I 5876A˚ and He II 4686A˚ emission features are indicated by black vertical dotted lines, while the feature near 6600A˚ which appears in the 2015 September 4 spectrum (see §4.2.2) is indicated by a vertical red dotted line. The transient spectra show many broad emission features and blue continuum emission in early epochs which rapidly fade, with later spectra resembling that of the host galaxy rather than that of the transient.

97 % − #" & # − − #" 01., # − ' − 06 $ − #" ( − #" +0474360-2 ) − #"

"!' # $ ' *3.5/80:1.,

Fig. 4.4.— Unfolded X-ray spectrum (unbinned) of ASASSN-15oi ﬁtted with an +10 absorbed black body plus power law model, with a blackbody temperature of 49−9 +1.50 eV and a power law index of Γ = 1.76−0.84. The dotted lines display the black body and power law models and the solid line shows the combined model. The data were ﬁtted with Cash Statistics (Cash 1979).

98 11 Days 10.5 16 Days 24 Days 10.0 32 Days 43 Days 70 Days 9.5 100 Days He II )

⊙ X-ray 9.0 ) (L λ

λL 8.5

log( 8.0

7.5

7.0

10 100 1000 10000 Rest Wavelength (A)˚

Fig. 4.5.— Evolution of the SED of ASASSN-15oi shown in different colors (points) and the corresponding best-fitting blackbody models for each epoch (lines). All data have been corrected for Galactic extinction and host contribution. Error bars are shown but can be smaller than the data points. All fits shown were made assuming the temperature prior described in §4.2.1. The X-ray spectral model inferred from the average of all the X-ray data (see Figure 4.4 and the ionizing luminosity implied by the He II 4686A˚ line (see §4.2.2) are shown in black. Neither the X-ray emission nor the ionizing luminosity for the He II line can be explained by the single blackbody model inferred from the optical/UV data, and a more complex model is likely needed.

99 Fig. 4.6.— Evolution of ASASSN-15oi’s X-ray (red triangles) and optical/UV (black squares) luminosity over time assuming the emission is consistent with a blackbody with the temperature prior described in the text. Filled points indicate those epochs where 4 or more data points were used to fit the temperature, and unfilled points indicate those epochs where fewer data points were used. The dashed line shows the −5/3 best L ∝ (t − t0) power law fit with t0 = 57228.0, while the dotted line shows the −α best unconstrained power law fit L ∝ (t − t0) with t0 = 57212.3 and the power law index α = 1.62. The solid line shows the best exponential fit, L ∝ e−(t−t0)/τ , with t0 = 57241.6 and τ = 46.5 days. While all three models fit the data reasonably well, the exponential is the best match to the data.

100 1.0

-19 days PTF09ge 0.5

0 +7 days

) + constant -1 day

λ ASASSN-15oi SN 2004gq log(f

−0.5

+0 days SN 2002ap −1.0 4000 5000 6000 7000 8000 9000 Rest Wavelength (A)˚

Fig. 4.7.— Comparison between spectra taken at or before maximum of ASASSN-15oi (Prentice et al. 2015), the He-rich TDE PTF09ge (Arcavi et al. 2014), the Type Ib SN SN 2004gq (Modjaz et al. 2014), and the broad-line Type Ic SN SN 2002ap (Gal-Yam, Ofek & Shemmer 2002). The epoch of each spectrum in days-since maximum-light is given next to each spectrum. (The epoch of ASASSN-15oi is days-since-discovery, since the peak date is not known.) Spectroscopically, ASASSN-15oi is far more similar to PTF09ge than it is to either SN, which have diﬀerent emission and absorption features and diﬀerent continuum shapes.

101 Fig. 4.8.— Comparison of (UV M2 − UV W 1) (top panel) and (UV M2 − U) (bottom panel) color evolution between ASASSN-15oi (blue pentagons); ASASSN- 14ae (Holoien et al. 2014, black squares); and ASASSN-14li (Holoien et al. 2016a, red triangles). Extinction correction and host ﬂux subtraction has been applied to all objects. All three TDEs show little evolution in (UV M2 − UV W 1) and remain quite blue in both colors for months after discovery. ASASSN-15oi does become bluer in (UV M2−U) before leveling oﬀ in later epochs, indicating an increasing temperature.

102 ASASSN-14ae ASASSN-14li 10.5 ASASSN-15oi )

⊙ 10.0 ) (L λ λL 9.5 log(

9.0

1000 10000 Rest Wavelength (A)˚

Fig. 4.9.— Comparison of the blackbody SEDs inferred from early epoch optical/UV data for ASASSN-15oi (blue solid line, 11 days after discovery), ASASSN-14ae (black dashed line, days after discovery), and ASASSN-14li (red dash-dotted line, 8 days after discovery). In all three cases, the temperature was fit to host-subtracted and extinction-corrected Swift and ground photometric data, which are also shown on the figure in the same colors and symbols used in Figure 4.8. ASASSN-14li and ASASSN- 15oi were fit with a temperature prior. Though more luminous, ASASSN-15oi exhibits a blackbody temperature that is not uncommon for an optically-selected TDE.

103 Fig. 4.10.— Comparison of the evolution of the luminosity inferred from the blackbody SED ﬁts for ASASSN-15oi (blue pentagons), ASASSN-14ae (black squares), and ASASSN-14li (red triangles). The X-ray emission detected in ASASSN- 15oi and ASASSN-14li is not included in the luminosity measurements. The scale on the right shows the luminosity in units of the Eddington luminosity for a 7 MBH = 10 M⊙ black hole. ASASSN-15oi is the most luminous of the three TDEs by a clear margin, but its luminosity declines at a rate similar to that of ASASSN-14ae, and the slower-fading ASASSN-14li is of comparable luminosity roughly 70 days after discovery.

104 Fig. 4.11.— Comparison of the blackbody temperature evolution inferred for ASASSN-15oi (blue pentagons), ASASSN-14ae (black squares), and ASASSN-14li (red triangles). The temperature was fit with a prior for all epochs of ASASSN-15oi and ASASSN-14li and for those epochs of ASASSN-14ae that did not include Swift observations, as described in the text. While ASASSN-14ae and ASASSN-14li exhibit the constant temperature evolution common in TDEs, ASASSN-15oi shows significant evolution, becoming much hotter than both other ASAS-SN TDEs in the first ∼ 70 days following discovery, although the maximum temperature is prior-dependent.

105 Fig. 4.12.— Comparison of the evolution of the photospheric radius inferred from the blackbody SED ﬁt for ASASSN-15oi (blue pentagons), ASASSN-14ae (black squares), and ASASSN-14li (red triangles). The left scale gives the radius in units of the 7 gravitational radius of a MBH = 10 M⊙ black hole, and the right scale converts this 1/2 into a velocity as v0 = c(2rg/r) . For comparison, the horizontal line gives the tidal disruption radius RT of the Sun, and the dashed line shows the radial evolution r(t) of an orbit expanding at v0(r). The photospheric radii are roughly bounded by r(t) at peak and then shrink back towards the tidal radius. ASASSN-15oi shrinks fastest, going from the largest to the most compact in roughly 70 days.

106 Fig. 4.13.— Comparison of the evolution of the FWHM (right scale, km s−1) of prominent emission lines in each ASAS-SN TDE. The left scale shows the radius 1/2 evolution r(t) assuming a radial velocity of v0 = c(2rg/r) . Also as in Figure 4.12, the horizontal line gives the tidal disruption radius RT of the Sun, and the dashed line shows r(t). Results are shown for the most prominent emission line in each ASAS-SN TDE: Hα is shown for ASASSN-14ae (black squares) and ASASSN-14li (red triangles), while He II 4686A˚ is shown for ASASSN-15oi (blue pentagons). To the extent that the velocity can be mapped into a radius, the line emission regions lie outside the photospheric radii (Figure 4.12), close to or beyond r(t), and evolve to larger radii. ASASSN-15oi exhibits a much more rapid narrowing in its emission lines than ASASSN-14ae or ASASSN-14li.

107 Filter Magnitude Uncertainty

NUV > 20.57 — V 17.13 0.05 J 15.17 0.07 H 14.52 0.07

Ks 14.05 0.09 W 1 14.18 0.03 W 2 14.12 0.05

′′ Measured 5.0-radius aperture JHKs magnitudes from 2MASS and a 14′′.0-radius aperture V -band magnitude from ASAS- SN. The WISE W 1 and W 2 are PSF photometry magnitudes from the AllWISE source catalog. The GALEX NUV limit is a 3-sigma upper limit measured from co-added exposures totaling ∼ 200s of exposure time. All magnitudes are in the Vega system.

Table 4.1. Photometry of the Host Galaxy of ASASSN-15oi.

108 Filter Magnitude Uncertainty

UV W 2 21.54 0.13 UV M2 21.20 0.17 UV W 1 20.02 0.10 U 18.44 0.07

BUVOT 18.18 0.06

BLCOGT 18.16 0.06

VUVOT 17.28 0.05

VLCOGT 17.20 0.05 I 16.05 0.04

Estimated 5′′.0-radius aperture magnitudes of the host galaxy synthesized from the host galaxy SED ﬁt.

Table 4.2. Estimated Photometry of the Host Galaxy of ASASSN-15oi.

UT Date MJD Telescope/Instrument Exposure (s)

2015 September 04.22 57269.22 du Pont-2.5m/WFCCD 1 × 900 2015September 12.20 57277.20 MDM-2.4m/OSMOS 1 × 600 2015 October 11.82 57306.82 Tillinghast-1.5m/FAST 1 × 2400 2015 October 12.82 57307.82 Tillinghast-1.5m/FAST 1 × 2100 2015 October 15.75 57310.75 Tillinghast-1.5m/FAST 1 × 1320 2015October17.12 57312.12 duPont-2.5m/WFCCD 2 × 900 2015 November 07.06 57333.06 du Pont-2.5m/WFCCD 2 × 1200

Table 4.3. Spectroscopic Observations of ASASSN-15oi.

109 Chapter 5: The Unusually Luminous Type-Defying II-P/II-L Supernova ASASSN-13co

In this chapter I describe the discovery and follow-up observations of ASASSN- 13co, an unusually luminous Type II SN with an atypically slow V -band decline rate, and the ﬁrst CCSN discovered by ASAS-SN.

Our transient source detection pipeline was triggered on 2013 August 29, detecting a new source with V = 16.9±0.1 mag at the coordinates RA = 21:40:38.72, Dec = +06:30:36.98 (J2000; Holoien et al. 2013). The object was also marginally detected on 2013 August 27 at roughly the same magnitude, but is not detected (V > 17 mag) in data obtained on 2013 August 23 and earlier. A search at the object’s position in the SDSS DR9 (Ahn et al. 2012) catalog revealed the source of the outburst to be the spiral galaxy PGC 067159 at redshift z = 0.023063, −1 −1 corresponding to a luminosity distance of d = 91.6 Mpc (H0 = 73 km s Mpc ,

ΩM =0.27, ΩΛ =0.73), and that the ASAS-SN source position was oﬀset by roughly 3′′.0 from the center of the host galaxy. Follow-up images obtained on 2013 August 30 by J. Brimacombe with a 16-inch RCOS telescope at Coral Towers Observatory (Cairns, Australia) and on 2013 September 3 with the Swift UVOT (Roming et al. 2005) conﬁrmed the detection of the transient. Figure 5.1 shows the archival SDSS g-band image of the host and an LCO (Brown et al. 2013) g-band image including the SN.

This chapter is adapted from “Discovery and Observations of the Unusually Luminous Type-

Defying II-P/II-L Supernova ASASSN-13co”, T. W.-S. Holoien, et al., Acta Astronomica, 66, 219,

(2016).

110 The archival SDSS spectrum of the host is that of a late-type star forming spiral galaxy. To obtain magnitudes of the host galaxy, we performed aperture photometry on archival SDSS ugriz images including all galaxy ﬂux and retrieved

2MASS JHKS magnitudes from the 2MASS Extended Source Catalog (Skrutskie et al. 2006). We then used Fast v1.0 (Kriek et al. 2009) to fit stellar population synthesis (SPS) models to the archival magnitudes of the host galaxy. The fit was made assuming an RV = 3.1 extinction law (Cardelli, Clayton & Mathis 1988), an exponentially declining star-formation history, a Salpeter IMF, and the 2 Bruzual & Charlot (2003) models. We obtained a good SPS fit (reduced χν =0.2), +0.3 10 +2.9 with the following parameters: M∗ = (1.7−0.7) × 10 M⊙, age= 7.1−4.7 Gyr, and +0.1 −1 SFR= 1.2−0.4 M⊙ yr .

A transient classiﬁcation spectrum obtained on 2013 September 1 with the Wide Field Reimaging CCD Camera (WFCCD) mounted on the Las Campanas Observatory du Pont 2.5-m telescope showed a blue continuum and Balmer lines with P-Cygni proﬁles characteristic of a young SN II. The broad Hα line showed a minimum P-Cygni absorption velocity of ∼ 12, 000 km s−1 and a possible detection of a faster component with a velocity of ∼ 19, 000 km s−1 (Morrell & Prieto 2013). With an absolute V -band magnitude of roughly −18.1 at detection, this was atypically luminous for an SN II, and we decided to start a follow-up campaign in order to fully characterize this interesting transient.

In §5.1 I describe data taken of the SN during our follow-up campaign. In §5.2 I analyze these data and describe the properties of the transient, using a recent phenomenological model (Pejcha & Prieto 2015a,b, henceforth jointly referred to as PP15) to ﬁt the SN’s multi-band light curves and derive its bolometric luminosity. Finally, in §5.3 I compare these properties to those of SNe in literature to examine the nature of the object.

5.1. Observations

In this section I summarize our photometric and spectroscopic observations of ASASSN-13co.

111 5.1.1. Photometric Observations

After detection of the transient, we were granted a series of Swift XRT (Burrows et al. 2005) and UVOT TOO observations between 2013 September 3 and 2013 October 19. We were also granted an additional late-time epoch of observations to be used for host flux subtraction on 2014 April 7. The Swift UVOT observations of ASASSN-13co were obtained in 6 filters: V (5468 A),˚ B (4392 A),˚ U (3465 A),˚ UV W 1 (2600 A),˚ UV M2 (2246 A),˚ and UV W 2 (1928 A)˚ (Poole et al. 2008). We used the UVOT software task Uvotsource to extract the source counts from a 5′′.0 radius region and a sky region with a radius of ∼40′′.0. The UVOT count rates were converted into magnitudes and fluxes based on the most recent UVOT calibration (Poole et al. 2008; Breeveld et al. 2010).

The XRT was operating in Photon Counting mode (Hill et al. 2004) during our observations. The data from all epochs were reduced and combined with the software tasks Xrtpipeline and Xselect to obtain an image in the 0.3−10 keV range with a total exposure time of ∼ 25, 300 s. We used a region with a radius of 20 pixels (47′′.1) centered on the source position to extract source counts and a source-free region with a radius of 100 pixels (235′′.7) for background counts. From this combined image, we detect a source at the position of the SN at a level of +2.1 −4 −1 5.0−1.8 × 10 counts s . To convert this to a flux, we assume a power law spectrum with Γ = 2 and Galactic H I column density (Kalberla et al. 2005), yielding a flux of ∼ 2.5 × 10−14 erg cm−2 s−1. At the host distance of d = 92 Mpc, this corresponds 40 −1 6 to a luminosity of LX = 2.5 × 10 erg s (6.6 × 10 L⊙). Due to a lack of prior X-ray data for the host galaxy in archival data from Chandra, XMM-Newton, or the ROSAT All-Sky Survey (Voges et al. 1999), we cannot determine whether this X-ray flux is from the SN or the host, and we do not include the X-ray detection in the analyses that follow.

In addition to the Swift observations, we obtained gri images with the LCO (Brown et al. 2013) 1-m telescopes located at Sutherland, South Africa, Cerro Tololo, Chile, and at MacDonald Observatory between 2013 October 4 and 2013 December 14. We also obtained an image subtraction template epoch in all three ﬁlters on 2014 June 5 from MacDonald Observatory. We measured aperture photometry on the

112 LCO images using a 5′′.0 aperture radius to match the Swift UVOT measurements. The photometric zero points were determined using several SDSS stars in the ﬁeld. All photometric data are shown in Figure 5.2.

Figure 5.2 shows the UV and optical light curves of ASASSN-13co from MJD 56528.1, the estimated explosion date from the PP15 fits (see §5.2), to our latest epoch of observations on MJD 56641.1 (113 days after explosion) without extinction correction or host flux subtraction. Also shown are the dates of spectroscopic follow-up observations and the 5′′.0 aperture magnitudes measured at the position of the SN in the late-time subtraction template images. The gri data at all epochs and UV data at later epochs are heavily contaminated by host galaxy flux.

In order to isolate the SN flux for more accurate photometric measurements, we performed host flux subtraction in two ways. For the LCO gri images, we aligned the 2014 June 5 template images to the SN images using stars in the field and used Hotpants1, an implementation of the Alard (2000) algorithm for image subtraction, to subtract the templates from the science images. We then performed photometry using a 5′′.0 aperture on the subtracted images to obtain host-flux subtracted measurements of the SN flux. For the Swift data, we were unable to use Hotpants because there are too few stars in the field. Instead, we measured the 5′′.0 host galaxy flux at the position of the SN in the 2014 April 7 template image. We then subtracted this flux from the flux measurements made in the science epochs, producing host-subtracted flux measurements that were then converted to host-subtracted magnitudes. These host-subtracted magnitudes were corrected for Galactic extinction and are used in the light curve fits and analysis presented in §5.2.

5.1.2. Spectroscopic Observations

We obtained nine low- and medium-resolution optical spectra of ASASSN-13co spanning more than two months between 2013 September 1 and 2013 November 15. The spectra were obtained with WFCCD mounted on the LCO du Pont 2.5-m telescope (range 3700 − 9500 A,˚ R ∼ 7 A),˚ the MDM Modular Spectrograph

1http://www.astro.washington.edu/users/becker/v2.0/hotpants.html

113 (Modspec) mounted on the MDM Observatory Hiltner 2.4-m telescope (range 4660 − 6730 A,˚ R ∼ 4 A),˚ the Ohio State Multi-Object Spectrograph (OSMOS; Martini et al. 2011) mounted on the MDM Observatory Hiltner 2.4-m telescope (range 4200 − 6800 A,˚ R ∼ 4 A),˚ and with DIS mounted on the Apache Point Observatory 3.5-m telescope (range 3500 − 9600 A,˚ R ∼ 7 A).˚ The spectra were reduced using standard techniques in Iraf. We applied telluric corrections to all the spectra using the spectrum of the spectrophotometric standard observed the same night. Figure 5.3 shows a montage of the ﬂux-calibrated spectra.

The main characteristics of the spectra of ASASSN-13co are a strong blue continuum in the early epochs that becomes weaker over time and broad emission features in all epochs, most notably Hα and Hβ, with no narrow emission lines that would indicate circumstellar interaction. These features are characteristic of a Type II SN, and we used the first follow-up spectrum of ASASSN-13co (from UT 2013 September 01) to classify the SN using the Supernova Identification code (Snid; Blondin & Tonry 2007). The best Snid matches were normal Type II-P SNe (SNe II-P) around maximum light and at a redshift z = 0.023, consistent with the host galaxy redshift measurement in Springob et al. (2005). Based on the SNID results and non-detection limits from ASAS-SN, we constrained the explosion date of ASASSN-13co to MJD 56529.1±1 day. This constraint is used in §5.2 when fitting the SN light curve with the PP15 model.

We compare the spectra of ASASSN-13co at three epochs to two SNe II that showed good matches with ASASSN-13co in SNID, SN 2006bp (Quimby et al. 2007) and SN 1999gi (Leonard et al. 2002) in Figure 5.4. All archival spectra of comparison objects were retrieved from the Weizmann Interactive Supernova data REPository (WISEREP; Yaron & Gal-Yam 2012). The comparison conﬁrms that, spectroscopically, ASASSN-13co resembles a normal SN II-P.

Finally, we used a number of emission lines in ASASSN-13co’s spectra (indicated in of Figure 5.3) to measure photospheric velocities at diﬀerent epochs. Figure 5.5 shows the velocity measured for each line in each epoch it was detected. The line velocity evolution is consistent with that of other SNe II (e.g., Quimby et al. 2007;

114 Valenti et al. 2014), providing further evidence that, spectroscopically, ASASSN-13co does not show any features that diﬀerentiate it from normal SNe II-P.

5.2. Light Curve Fits and Analysis

The PP15 model was developed in order to better understand the physics that drive the observed differences between SNe II-P and SNe II-L. This new phenomenological model uses multi-band photometric measurements and line velocities to derive the photospheric radius and temperature variations of an SN. These variations can then be used to calculate light curves in other filters, bolometric luminosities, Nickel-56 masses, and other quantities of interest. The model is based on a sample of 26 well-observed SNe II-P and is designed for fitting such SNe, but PP15 show that it works for SNe II-L as well. The PP15 model, combined with detailed observations of nearby SNe II, provides a new tool for investigating the diversity of these SNe and the physical processes that drive their observed properties. In particular, unusual events, such as particularly luminous SNe or those with light curve shapes that do not match typical Type II-P or Type II-L shapes, may be especially revealing. The PP15 model also provides a method of generating light curves in additional filters beyond those that were used for follow-up observations of SNe II, allowing the comparison of SNe with disparate datasets.

Using the host-subtracted photometry and spectroscopic Fe II (5169 A)˚ line expansion velocities we fit the multi-band light curves of ASASSN-13co using the PP15 model. When fitting the light curves of ASASSN-13co, we constrained the explosion date to MJD 56529.1±1.0 day so that the fits will be consistent with our spectroscopic observations. We performed the fit with the distance modulus unconstrained, finding a best-fit value of 34.46 ± 0.23, which is roughly 0.4 magnitudes off from the value of 34.81 obtained assuming the redshift of 0.023063 from (Springob et al. 2005). However, the small number of expansion velocities and the photometric data for ASASSN-13co indicate that the distance determined using the expanding photosphere method implemented in PP15 might be unreliable. Therefore, we also fit the light curve with a fixed distance modulus of 34.81. There are no noticeable differences in the light curve fits, and we use these results, including

115 the best-fit explosion date of MJD 56528.1, in the analyses that follow. Figure 5.6 shows the host-subtracted photometry and the light curve fits from the PP15 model. The best-fit parameters from the PP15 model are given in Table 5.2.

The shape of ASASSN-13co’s light curve appears to resemble that of an SN II-L more than that of an SN II-P. However, the derived model parameters shown in Table 5.2 do not seem to fall nicely into either category. The SNe II-P in the PP15 sample typically showed plateau durations of 85 days ∼< tp ∼< 130 days and transition widths of 1 day ∼< tw ∼< 6 days, while the SN II-L SN 1980K had a plateau duration of tp ≈ 61 days and a transition width of tw ≈ 13 days. ASASSN-13co appears to fall somewhere between these two classes, showing both a fairly long plateau phase (though one still shorter than the smallest for the II-P sample) and a long transition width, longer even than that of SN 1980K. We note that the exact values of both the plateau duration and transition width are not well-constrained due to the lack of later photometric data, but the best-ﬁt values are consistent with the light curve shape. This seems to indicate that ASASSN-13co is neither a Type II-P nor a Type II-L SN. As discussed later in this section, this may further illustrate that SNe II exhibit a continuum of morphologies rather than two distinct types.

We also obtain an estimate of the evolution of the bolometric luminosity of ASASSN-13co from the PP15 model, which we compare to the SN sample from PP15 in Figure 5.7. The PP15 model constructs the bolometric curve of the input SN by calculating reddening-corrected spectral energy distributions (SEDs) at many epochs from the derived multi-band light curves, and then integrating the flux across all wavelengths in these SEDs using the trapezoidal rule. They cut off the integral at wavelengths shorter than 0.19 µm and estimate the long wavelength flux by extrapolating the K-band flux to infinity assuming the emission follows the Raleigh-Jeans tail of a blackbody. In Figure 5.7 we truncate the luminosity evolution of ASASSN-13co at the date of our last photometric epoch to avoid extrapolation to epochs that are not constrained by data. As can be seen in Figure 5.7, ASASSN-13co is more luminous than all but two of the SNe in the PP15 sample at early times, and its luminosity decreases steadily, rather than showing the rapid fall and “leveling off” characteristic of the SNe II-P in the sample. This

116 again indicates that ASASSN-13co’s light curve morphology does not resemble that of an SN II-P, despite the fact that its strongest spectral matches from SNID were all SNe II-P. In addition, Figure 5.7 shows that ASASSN-13co had an atypically high luminosity at maximum, and it could be even higher at earlier epochs, when emission in the UV was strongest and there might have been signiﬁcant emission at wavelengths shorter than the 0.19 µm cutoﬀ used in the luminosity integral.

In order to check the luminosity fit for consistency, we use the (g − r) color of ASASSN-13co at 41 days past explosion (the first date for which we have LCO photometry) to calculate a g-band bolometric correction of BCg = −0.63 ± 0.09 based on Lyman, Bersier & James (2014). Assuming E(B − V )=0.05, we then calculate a bolometric magnitude from the g-band magnitude and the bolometric correction and convert this to a bolometric luminosity, obtaining 8 Lbol(texp + 41 days)=(8.6 ± 0.9) × 10 L⊙. The bolometric luminosity fit by the PP15 9 model at this time is Lbol(texp + 41 days)=(1.4 ± 0.2) × 10 L⊙, consistent with the Lyman, Bersier & James (2014) value to roughly 2 sigma. This difference is not unexpected, as the PP15 model overestimates the g-band magnitude in its fit for ASASSN-13co (see Figure 5.6). If we use the value of E(B − V )=0.28 that is found 9 by the PP15 model, we obtain Lbol(texp + 41 days)=(1.3 ± 0.9) × 10 L⊙ using the Lyman, Bersier & James (2014) method, indicating that the PP15 fit for E(B − V ) is perhaps more consistent with the data.

Recently there has been some debate about whether SNe II truly fall into distinct “II-L” and “II-P” groups (e.g., Arcavi et al. 2012; Faran et al. 2014a,b) or whether SN II light curves can have a variety of shapes that do not necessarily conform to one of the two traditional types (e.g., Anderson et al. 2014; Sanders et al. 2015). The distinction may be more apparent in some photometric bands than in others; in particular, Arcavi et al. (2012) use R-band light curves while Anderson et al. (2014) use V -band. In Figure 5.8 we compare the absolute V -band light curve of ASASSN-13co, derived from the PP15 model ﬁt, to those of the SN II sample used in Figure 2 of Anderson et al. (2014), and in Figure 5.9 we compare the absolute I-band light curve of ASASSN-13co to those of the Optical Gravitational Lensing Experiment (OGLE) Type II SN sample from Poznanski et al. (2015).

117 The two Figures illustrate that the II-P or II-L classiﬁcation can be highly dependent on the photometric band used to observe the SN. The V -band light curve shown in Figure 5.8 does not seem to be consistent with any of the others in the comparison group: it has a higher absolute magnitude at peak, and although it does appear to decline steadily, like an SN II-L, it appears to do so at an atypically slow rate. While it does not show a distinct “plateau”, the slow decline appears to be unusual, particularly for an SN II as luminous in the V -band as ASASSN-13co.

Conversely, the I-band light curve of ASASSN-13co appears to be largely consistent with the OGLE Type II SN sample shown in Figure 5.9. To obtain the OGLE light curve fits shown in the Figure, we performed spline fits to the photometric data from Poznanski et al. (2015), while the I-band fit for ASASSN-13co was again derived from the PP15 model. While ASASSN-13co remains one of the more luminous SNe in the sample, there are other SNe with similar I-band light curve shapes, and the trend of more luminous SNe declining more rapidly is less apparent in the I-band sample than it is in the V -band sample. Based on its steady decline rate, ASASSN-13co likely would have been classified as an unusually luminous SN II-L based on its I-band light curve. That aside, the OGLE sample is composed of only 11 SNe, and appears to be incomplete at fainter magnitudes. A larger sample spanning a wider magnitude range may well exhibit a more continuous range of decline rates.

Anderson et al. (2014) found a correlation between the peak absolute V -band magnitude and the slope of the V -band light curve’s decline during the plateau phase, indicating that SNe with smaller peak magnitudes tend to fade at a faster rate than those with higher peak magnitudes. PP15 were able to reproduce this trend using the absolute V -band magnitude and light curve slope at constant eﬀective temperature at τ = 0, corresponding to roughly the middle of the plateau phase, rather than the peak magnitude and “plateau phase” slope used by Anderson et al. (2014). In Figure 5.10 we reproduce Figure 9 from PP15, showing the absolute magnitudes and slopes of the SNe in their sample, along with ASASSN-13co. Figure 5.10 conﬁrms what can be seen by eye in Figure 5.8—while ASASSN-13co is

118 one of the most luminous SNe in Figure 5.8, its V -band decline slope seems to be more consistent with SNe that are two or more magnitudes fainter.

Ultimately, ASASSN-13co appears to be photometrically unusual. It is quite luminous for an SN II, with a peak V -band absolute magnitude of ∼ −18.1. Its light curve shape and PP15 model parameters indicate that while its I-band decline rate is fairly typical of an SN II-L, its V -band properties fall somewhere between those of a typical SN II-L and those of an SN II-P, perhaps lending credence to the idea that SNe II should not be divided into two distinct groups. However, it also does not conform to the magnitude-slope relations derived in Anderson et al. (2014) or PP15, as its light curve declines at a rate more typical of an SN that is ∼ 2 magnitudes fainter. While it is spectroscopically very similar to many SNe II-P, ASASSN-13co’s photometric properties are inconsistent with that designation.

5.3. Discussion

Our observations indicate that ASASSN-13co was an unusual event. Its spectra match well to normal SNe II-P, but its light curve shape is a better match to that of an SN II-L, and its peak magnitude (MV ∼ −18.1) is quite high. Light curve fits with the PP15 model corroborate this picture, showing that the bolometric luminosity was atypically high and that its light curve declined at a slow but steady rate. Because it does not fit into either of the traditional designations, we conclude that ASASSN-13co provides additional evidence that SNe II exhibit a wide range of physical and observed properties that cannot be quantified by the simple distinction between II-P and II-L. However, this distinction may also depend on the photometric bands used for follow-up observation. More observational evidence in multiple photometric bands will be required to definitively determine whether the II-P/II-L distinction is real, or whether SNe II simply exhibit a broad but continuous range of photometric properties.

ASASSN-13co has proven to be an example of the usefulness of both the PP15 model and the ASAS-SN project. While the PP15 model was designed speciﬁcally to model the light curves of SNe II-P, it is able to provide a reasonable ﬁt to our data

119 despite ASASSN-13co exhibiting a light curve that declines at an unusually slow rate. That we are able to extract a bolometric luminosity curve and other properties from these fits that do not depend on theoretical models provides an example of how the PP15 model will be a powerful tool for investigating the diverse properties of SNe II in the future. It also allows us to compare objects with different photometric datasets, as we have done in our comparison with the Anderson et al. (2014) and Poznanski et al. (2015) samples: ASASSN-13co has no Johnson V - or I-band data, but the PP15 model allows us to “predict” that flux, making a comparison possible.

ASASSN-13co represents a similar success for the ASAS-SN project, as it is an unusual event and provides an example of the quality of study that can be done with ASAS-SN discoveries. As ASAS-SN comprises only small telescopes, its discoveries can be observed in great detail using relatively small 1- and 2-m class telescopes, and with roughly 500 SN discoveries (including 90 SNe II) at time of writing, it will undoubtedly be an extremely useful project for investigating the diversity of these events in the future.

120 Fig. 5.1.— Discovery image of ASASSN-13co. The top-left panel shows the ASAS- SN V -band reference image and the top-right panel shows the ASAS-SN subtracted image from 2014 August 29, the date of discovery. The bottom-left panel shows the archival SDSS g-band image of the host galaxy PGC 067159 and the right panel shows an LCO g-band image taken during the SN. The dates of the observations are listed on each panel. The red circles have radii of 2′′.0 and are centered on the position of the SN in the LCO image.

121 UVW2−2.0 g+2.1 UVM2−1.2 V+3.1 15 UVW1−0.5 r+4.2 U+0.3 i+5.3 B+1.2 16

18 constant +

mag 19

0 20 40 60 80 100 120

Days Since Explosion

Fig. 5.2.— Light curves of ASASSN-13co, starting at the estimated explosion date (MJD = 56528.1) and spanning 113 days. Follow-up data obtained from Swift (UV + optical) and the LCO 1-m (optical) telescopes are shown as circles. All magnitudes are in the Vega system. The data are not corrected for extinction and error bars are shown for all points, but in some cases they are smaller than the data points. Host galaxy magnitudes measured in a 5′′.0 aperture centered on the position of the SN in the subtraction template images are shown as stars at −5 days. Dates of spectroscopic follow-up are indicated with vertical black bars at the bottom of the ﬁgure. The measurements are heavily contaminated by host galaxy ﬂux in the optical bands.

122 Hβ Fe II He I Hα OI 10 Si II 2013/09/01

2013/09/02

9 2013/09/03 2013/09/04

2013/09/05 8 2013/09/11 ) + constant λ 2013/09/17 log(f

2013/10/26 7

2013/11/15

4000 5000 6000 7000 8000 Rest Wavelength (A)˚

Fig. 5.3.— Spectral time-sequence of ASASSN-13co. Each spectrum is labelled by the UT date on which it was obtained. Emission lines used to measure velocities in Figure 5.5 are labelled and marked with vertical dashed lines. The rapidly fading blue continuum and broad hydrogen emission features are typical of a Type II SN.

123 11 SN 2006bp ASASSN-13co SN 1999gi

10 +8 days

+8 days +4 days

9 +25 days +24 days

8 +27 days ) + constant λ

+57 days log(f 7

+63 days

+85 days 6

4000 5000 6000 7000 8000 9000 Rest Wavelength (A)˚

Fig. 5.4.— Spectroscopic comparisons between ASASSN-13co and Type II-P SNe, SN 2006bp (Quimby et al. 2007) and SN 1999gi (Leonard et al. 2002) at three epochs. The evolution of ASASSN-13co looks very similar to both comparison objects, particularly SN 1999gi.

124 Fig. 5.5.— Velocity evolution of the emission lines used to measure velocities. Each line is shown with a diﬀerent symbol and color. Symbols represent the measurements, while the connecting lines are provided only to guide the eye. The velocities are higher than typical for a Type II SN, but the evolution is normal. Table 5.1 contains all the measured velocity data.

125 UBV R IJH K g r i z 16 16

17 17

18 18

19 19 magnitude magnitude

20 20

21 21 10000 Swift uvw2 uvm2 uvw1 u b v ] 1 16 − 8000

6000 18

4000 19 magnitude

20 2000

Fe II expansion velocity [km s 21 0 56550 56600 56550 56600 JD − 2 400 000 JD − 2 400 000

Fig. 5.6.— Host-subtracted light curves and Fe II velocities of ASASSN-13co in various filters from the PP15 model fits (colored lines) and the corresponding host- subtracted photometry and velocity measurements (colored points) covering roughly 3.5 months after the explosion (MJD 56528.1, shortly after the final ASAS-SN non- detection on MJD 56527.9). 3-sigma upper limits are shown as downward-pointing triangles. Error bars are shown for all points. Filters are indicated in each panel with the color of the text matching that of the corresponding light curve. The model seems to fit all bands fairly well despite a lack of wavelength coverage, and we truncate the fits shortly after our last epoch of photometric data, as later epochs are poorly constrained by the model. Based on these fits, ASASSN-13co appears to resemble an SN II-L more than an SN II-P, though it declines more slowly than a typical SN II-L. Table A.6 contains all the host-subtracted photometric data.

126 Fig. 5.7.— Evolution of the bolometric luminosity of ASASSN-13co (red) compared to the SNe used in Figure 13 of PP15 (grey). The 1−σ uncertainty region is shown in red around the curve. We truncate the luminosity curve for ASASSN-13co at the time of our ﬁnal epoch of photometric data to avoid extrapolation by the model. At early times in particular, ASASSN-13co is one of the most luminous SNe of those shown, and its luminosity decline does not show a strong plateau phase.

127 ASASSN-13co Anderson et al. (2014) Sample −18

−17

−16

−15 Absolute V-band magnitude

−14

−13 0 20 40 60 80 100 120

Time past explosion epoch (days)

Fig. 5.8.— Comparison of the absolute V -band light curve of ASASSN-13co (red) to those of the SN II sample shown in Figure 2 of (Anderson et al. 2014) (grey). The light curve for ASASSN-13co is derived from the PP15 model ﬁt and assumes a distance modulus of 34.81. All light curves shown are corrected for Galactic extinction but not for host galaxy extinction. ASASSN-13co has one of the most luminous V -band light curves of all the SNe shown, and appears to have a shallower decline than many of the luminous SNe in the (Anderson et al. 2014) sample.

128 −19.0 OGLE ASASSN-13co −18.5

−18.0

−17.5

−17.0

−16.5

−16.0 Absolute I-band magnitude

−15.5 0 20 40 60 80 100

Time past explosion epoch (days)

Fig. 5.9.— Comparison of the absolute I-band light curve of ASASSN-13co (red) to those of the OGLE SN II sample from Poznanski et al. (2015) (grey). The light curve for ASASSN-13co is derived from the PP15 model fit and assumes a distance modulus of 34.81. The OGLE light curve fits are spline fits to the I-band photometric data. All light curves shown are corrected for Galactic extinction but not for host galaxy extinction. ASASSN-13co has one of the most luminous I-band light curves of all the SNe shown, but its decline rate does not seem atypical from the rest of the sample.

129 0.025 ASASSN-13co PP15 Sample 0.020 ] 1 − 0.015

[mag day 0.010 = 0 τ

0.005 slope at

0.000

−0.005 −14 −15 −16 −17 −18 MV (τ = 0)

Fig. 5.10.— Correlation of the absolute V -band magnitude and the V -band light curve slope measured at τ = 0 for ASASSN-13co (red) and the SN II sample shown in Figure 9 of PP15 (black). ASASSN-13co has a shallower slope than other SNe with similar magnitudes, conﬁrming what can be seen by eye in Figure 5.8.

130 Hβ Fe II He I Si II Hα O I MJD (4861 A)˚ (5169 A)˚ (5875 A)˚ (6355 A)˚ 6563 A)˚ (7773 A)˚

56536.10 −11478 — −11220 −9562 −11603 — 56537.14 −11576 — — — — — 56538.10 −11040 — −11230 −9468 −11863 — 56539.18 −10831 — −11399 −9765 −11297 — 56540.14 −11589 −9356 −11159 −8930 −11219 — 56546.26 −10269 −8885 −10195 −9142 −11114 — 56552.29 −9647 −8615 −9794 −8270 −11022 — 56591.22 −7273 −5455 −6528 — −8544 −6213 56611.08 — −4599 −6128 — −8153 —

Column headers indicate the lines and wavelengths used to measure the velocities. All velocities are listed in km s−1 and no value is given for epochs where a measurement was not possible.

Table 5.1. Photospheric Velocities of ASASSN-13co.

131 Parameter Value Uncertainty Rescaled

Explosion time (texp) 56528.1 — —

Plateau duration (tp) 84.650 3.674 5.922

Transition width (tw) 26.263 1.450 2.576 3 −1 ω0 (10 km s ) 124 11 14

ω1 −0.809 0.033 0.052 −1 ω2 (km s ) 0.0 — —

γ0 −0.0044 0.0005 0.0006

γ1 11.30 0.06 0.11

α0 −0.0011 0.0002 0.0004

α1 0.12 0.00 0.01 Distance modulus 34.81 0.00 0.00 E(B − V ) 0.28 0.03 0.05

Variable names correspond to those deﬁned in PP15. The “Rescaled” values are the uncertainties scaled so that χ2/DOF=1, as described in PP15. An uncertainty of 0.0 indicates that the parameter’s value was

ﬁxed. In the cases of texp and ω2, the model converged on the lower bound of the allowed range for the parameter, and the uncertainty is not calculated. Given the photometric data constraints, the uncertainty

on texp is about 2 days, but this uncertainty is not propagated into the other parameters.

Table 5.2. PP15 Model Parameters for ASASSN-13co.

132 Chapter 6: The ASAS-SN Bright Supernova Catalogs

Systematic searches for SNe have a long and venerable history, beginning with the pioneering effort at Palomar by Zwicky (Zwicky 1938, 1942). In the modern era, the SN search effort has progressed through numerous survey projects which used varying degrees of automation to survey some or all of the sky for SNe and other transients, including the Lick Observatory SN Search (LOSS; Li et al. 2000), the Panoramic Survey Telescope & Rapid Response System (Pan-STARRS; Kaiser et al. 2002), the Texas SN Search (Quimby 2006), the SDSS SN Survey (Frieman et al. 2008), the Catalina Real-Time Transient Survey (CRTS; Drake et al. 2009), the CHilean Automatic SN sEarch (CHASE; Pignata et al. 2009), the Palomar Transient Factory (PTF; Law et al. 2009), the Gaia transient survey (Hodgkin et al. 2013), the La Silla-QUEST (LSQ) Low Redshift SN Survey (Baltay et al. 2013), the Mobile Astronomical System of TElescope Robots (MASTER; Gorbovskoy et al. 2013) survey, the Optical Gravitational Lensing Experiment-IV (OGLE-IV; Wyrzykowski et al. 2014), and the Asteroid Terrestrial-impact Last Alert System (ATLAS; Tonry 2011), among numerous others. However, despite the number of such surveys, there was no optical survey that surveyed the entire visible night sky on a rapid cadence to find the bright, nearby SNe that can be studied in the greatest detail until the creation of ASAS-SN in 2013.

This chapter is adapted from “The ASAS-SN Bright Supernova Catalog – I. 2013−2014”, T.

W.-S. Holoien, et al., MNRAS, 464, 2672, (2017), “The ASAS-SN Bright Supernova Catalog – II.

2015”, T. W.-S. Holoien, et al., MNRAS, 467, 1098, and “The ASAS-SN Bright Supernova Catalog

– III. 2016”, T. W-S. Holoien, et al., arXiv:1704.02320.

133 While ASAS-SN discovers fewer SNe overall than some other professional surveys, by design, all of ASAS-SN’s discoveries are bright and nearby, allowing them to be followed up over a wide wavelength range using only modest resources. A 1-m telescope is often more than sufficient to obtain a spectrum of an ASAS-SN discovery, and every ASAS-SN SN has been spectroscopically observed, confirmed, and classified. This is part because ASAS-SN data are announced publicly upon confirmation, allowing for rapid discovery and response by both the ASAS-SN team and others. Furthermore, because ASAS-SN’s survey approach is untargeted, our discoveries are not limited to specific types of galaxies. In fact, roughly a quarter of the host galaxies of ASAS-SN SNe have not had a previously determined spectroscopic redshift prior to the discovery of the SN, and in a few cases ASAS-SN hosts have not been identified as galaxies in any existing catalog. The ASAS-SN sample thus provides a new and unbiased tool for doing population studies of SNe and their host galaxies in the nearby universe.

In this chapter I present details and statistical analyses of all SNe discovered by ASAS-SN during its ﬁrst three-and-a-half years of operations, spanning 2013, 2014,

2015, and 2016. I also present the same information for all other bright (mV ≤ 17), spectroscopically conﬁrmed SNe discovered from 2014 May 1 through the end of 2016, providing a comparison to the ASAS-SN sample starting from the point where ASAS-SN became operational in both hemispheres. The cumulative combined sample now totals 668 SNe discovered since 2014 May 1, and we provide statistical analyses of these bright SNe and their host galaxies.

In §6.1 I describe the sources of the information used for these analyses. In §6.2, I give statistics on the SN and host galaxy populations in our full three-year cumulative bright SN sample, including the discoveries listed in Holoien et al. (2017a,b) and Holoien et al. (2017c), and discuss overall trends in the sample. Throughout my analyses, I assume a standard ΛCDM cosmology with −1 −1 H0 = 69.3 km s Mpc , ΩM = 0.29, and ΩΛ = 0.71 for converting host redshifts into distances. In §6.3, I report details of a retrospective examination of bright SNe from 2014 and 2015 that were not discovered or recovered by ASAS-SN in order to better understand the reasons why ASAS-SN would not ﬁnd SNe that are bright

134 enough to be detected. In §6.4, I conclude with our overall ﬁndings and discuss how the upcoming expansion to ASAS-SN will impact our future discoveries.

6.1. Data Samples

Below I outline the sources of the data collected in our SN and host galaxy samples. These data are presented in full in Holoien et al. (2017a), Holoien et al. (2017b), and Holoien et al. (2017c).

6.1.1. The ASAS-SN Supernova Sample

All names, discovery dates, and host galaxy names for ASAS-SN discoveries were taken from our discovery Astronomer’s Telegrams (ATels). Using archival classification and late-time spectra of the ASAS-SN SN discoveries taken from the Transient Name Server (TNS1) and Weizmann Interactive SN data REPository (WISEREP; Yaron & Gal-Yam 2012), we confirmed and updated the redshifts and classifications reported in classification telegrams. For those cases where an SN host had a previously measured redshift and the host redshift is consistent with the transient redshift, we list the redshift of the host taken from the NASA/IPAC Extragalactic Database (NED)2. For all other cases we used the redshifts measured from classification spectra. When available, we also collected the approximate age of the SNe at discovery measured in days relative to peak. Classifications were obtained using either Snid (Blondin & Tonry 2007) or the Generic Classification Tool (Gelato3; Harutyunyan et al. 2008), which both compare observed input spectra to template spectra in order to estimate the SN age and type.

Using the astrometry.net (Barron et al. 2008; Lang et al. 2010) package we solved the astrometry in follow-up images for all ASAS-SN SNe and measured a centroid position for the SN using Iraf. This approach typically yields errors of <1′′.0 in position, which is signiﬁcantly more accurate than measuring the SN

1https://wis-tns.weizmann.ac.il/ 2https://ned.ipac.caltech.edu/ 3gelato.tng.iac.es

135 position directly in ASAS-SN images, which have a 7′′.0 pixel scale. The images used to measure astrometry were obtained using the LCO 1-m telescopes (Brown et al. 2013), OSMOS (Martini et al. 2011) mounted on the MDM Hiltner 2.4-m telescope, or from amateur collaborators working with the ASAS-SN team. We used these coordinates and host coordinates available in NED to calculate the oﬀset from the host galaxy nucleus for each SN.

We re-measured V -band, host-subtracted discovery and peak magnitudes from ASAS-SN data for all ASAS-SN SN discoveries. We define the “discovery magnitude” as the magnitude of the SN on the announced discovery date. For cases with enough detections in the light curve, we also perform a parabolic fit to the light curve and estimate a peak magnitude based on the fit. We then used the brighter value between the brightest measured magnitude and the peak of the parabolic fit as the “peak magnitude” of the SN.

We collected information for all SNe discovered by ASAS-SN in 2013, 2014,

2015, and 2016, including those that peaked at magnitudes fainter than mV = 17. In the comparison analyses presented in §6.2, however, we only include those ASAS-SN

SNe with mV,peak ≤ 17 so that our sample is consistent with the non-ASAS-SN sample.

6.1.2. The Non-ASAS-SN Supernova Sample

We also collected information for all spectroscopically conﬁrmed SNe with peak magnitudes of mpeak ≤ 17 that were discovered by other professional and amateur SN searches between 2014 May 1 and 2016 December 31. These dates were chosen so that the sample could be compared to the ASAS-SN SNe discovered after ASAS-SN became operational in both hemispheres.

We compiled data for the non-ASAS-SN discoveries from the “latest supernovae” website4 designed and maintained by D. W. Bishop (Gal-Yam et al. 2013). This site compiles discoveries reported from diﬀerent channels and links objects reported

4http://www.rochesterastronomy.org/snimages/

136 by different sources at different times, making it an ideal source for collecting information on SNe discovered by different search groups. While we did use TNS for verification of the data from the latest supernovae website for discoveries made after 2015, we did not use it as the primary source of information on non-ASAS-SN discoveries, as some SN searches do not participate in the TNS system.

Names, discovery dates, coordinates, host names, host offsets, peak magnitudes, spectroscopic types, and discovery sources for each SN in the non-ASAS-SN sample were taken from the latest supernovae website when possible. Host galaxy redshifts were collected from NED when available and were taken from the latest supernovae website otherwise. For cases where a host name or host offset was not listed on the website for an SN, the primary name and offset were taken from NED. In these cases, we define the offset as the distance between the reported coordinates of the SN and the galaxy coordinates in NED. In some cases, no catalogued galaxy was listed at the position of the host in NED, but a host galaxy was clearly visible in archival Pan-STARRS data (Chambers et al. 2016) . In these cases, we measured the centroid position of the host nucleus using Iraf and calculated the offset using those coordinates. For all SNe in both samples we use the primary name of the host galaxy listed in NED, which sometimes differs from the name listed on the ASAS-SN SN page or the latest SN website.

We also used archival classiﬁcation spectra from TNS and WISEREP of several SNe discovered by non-ASAS-SN sources to update the redshifts and classiﬁcations of sources that had missing or incorrect information on the latest supernovae website.

For the purposes of the analyses presented here, I consider SNe discovered by non-professional astronomers as discovered by “amateurs”, and SNe discovered by all other professional searches as discovered by “other professionals.” In all 2.5 years of data collected, amateurs account for the largest number of bright SN discoveries after ASAS-SN.

We also noted whether or not the ASAS-SN team independently recovered these SNe while scanning our data. This is done to help quantify the impact ASAS-SN has on the discovery of bright SNe in the absence of other SN searches.

137 6.1.3. The Host Galaxy Samples

For the host galaxies of both SN samples, we collected Galactic extinction estimates for the direction to the host and host magnitudes in various photometric ﬁlters spanning from the near-ultraviolet (NUV) to the infrared (IR) wavelengths. Galactic

AV values from Schlaﬂy & Finkbeiner (2011) at the positions of the SNe were gathered from NED. NUV magnitudes were taken from the Galaxy Evolution Explorer (GALEX; Morrissey et al. 2007) All Sky Imaging Survey (AIS), optical ugriz magnitudes were gathered from the SDSS Data Release 12 (SDSS DR12; Alam et al. 2015) and Data Release 13 (SDSS DR13; Albareti et al. 2016), NIR JHKS magnitudes were gathered from 2MASS (Skrutskie et al. 2006), and IR W 1 and W 2 magnitudes were gathered from the WISE (Wright et al. 2010) AllWISE source catalog.

When a host galaxy was not detected in 2MASS, we adopted an upper limit corresponding to the faintest 2MASS host magnitudes in our sample for the J and H bands (mJ > 16.5, mH > 15.7). For hosts that are not detected in 2MASS but are detected in the WISE W 1 band, we estimated a host magnitude by adding the mean

KS − W 1 offset from the sample to the WISE W 1 data. This offset was calculated by averaging the offsets for all hosts that are detected in both the KS and W 1 bands from both SN samples from 2014 May 1 through 2016 December 31. The average offset is equal to −0.51 magnitudes with a scatter of 0.04 magnitudes and a standard error of 0.002 magnitudes. If a host was not detected in either 2MASS or WISE, we adopted an upper limit of mKS > 15.6, corresponding to the faintest detected host in our sample.

6.2. Analysis of the Sample

Combining all the bright SNe discovered between 2014 May 01, when ASAS-SN became operational in both hemispheres, and 2016 December 31 provides a sample of 668 SNe once we exclude ASAS-SN discoveries with mpeak > 17.0 (Holoien et al. 2017a,b,c). Of these, 58% (389) were discovered by ASAS-SN, 21% (137) were discovered by other professional surveys, and 21% (142) were discovered by amateur

138 astronomers. 449 were Type Ia SNe, 178 were Type II SNe, 40 were Type Ib/Ic SNe (SNe Ib/Ic), and 1 was a superluminous SN (SLSN). We consider Type IIb SNe as part of the Type II sample to allow for more direct comparison with the results of Li et al. (2011). ASASSN-15lh is excluded in analyses that follow looking at trends by type, as all available evidence points to it being an extremely luminous Type I SLSN (Dong et al. 2016; Godoy-Rivera et al. 2017), though it has also been classiﬁed as a tidal disruption event around a Kerr black hole (Leloudas et al. 2016). ASAS-SN discoveries account for 66% of the SNe Ia, 45% of the SNe II, and 25% of the SNe Ib/Ic. Amateur discoveries account for 16%, 30%, and 48% of the Type Ia, Type II, and Type Ib/Ic SNe in the sample, respectively, and discoveries from other professional surveys account for the remaining 18%, 25%, and 28% of each type.

Figures 6.1, 6.2, and 6.3 show pie charts breaking down the type distributions of SNe in the ASAS-SN, non-ASAS-SN, and combined samples, respectively. SNe represent the largest fraction of SNe in all three samples, as expected for a magnitude-limited sample (e.g., Li et al. 2011). Comparing to the “ideal magnitude-limited sample” breakdown predicted from the LOSS sample in Li et al. (2011), where there are 79% Type Ia, 17% Type II, and 4% Type Ib/Ic, the ASAS-SN sample matches the LOSS prediction almost exactly. The non-ASAS-SN sample and the combined sample have higher fractions of CCSNe, particularly SNe Ib/Ic.

ASAS-SN has been the dominant source of bright SN discoveries since it became operational in both hemispheres, and we often discover SNe shortly after explosion due to our rapid cadence: of the 336 ASAS-SN discoveries with approximate discovery ages, 69% (232) were discovered prior to reaching their peak brightness. ASAS-SN is also less aﬀected by host galaxy selection eﬀects than other bright SN searches. For example, 25% (96) of the ASAS-SN bright SNe were found in catalogued hosts that did not have previous redshift measurements available in NED, and an additional 4% (14) were discovered in uncatalogued hosts or have no apparent host galaxy. Conversely, only 16% (44) of non-ASAS-SN discoveries were found in catalogued hosts without redshift measurements, and only 3% (8) were in uncatalogued galaxies or were hostless.

139 The difference between ASAS-SN discoveries and SNe discovered by other searches is very clear when looking at their host galaxy luminosities and their offsets from their host nuclei. ASAS-SN discoveries have a smaller average offset from their host galaxy nuclei than bright SNe discovered by other searches. The host galaxy

KS-band absolute magnitudes and the offsets of the SNe from the host centers for all SNe in our sample are shown in Figure 6.4 (offset in arcseconds) and Figure 6.5 (offset in kiloparsecs). The median offsets and magnitudes are shown with horizontal and vertical lines for each SN source (ASAS-SN, amateurs, or other professionals). A luminosity scale corresponding to the magnitude scale is given on the upper axis of the figure to help put the magnitude scale in perspective, assuming that a typical

L⋆ galaxy has M⋆,KS = −24.2 (Kochanek et al. 2001).

Amateur surveys are significantly biased towards luminous galaxies and larger offsets from the host nucleus, which is unsurprising given that they tend to observe bright, nearby galaxies and use less sophisticated detection techniques than professional surveys. This approach allows amateur observers to obtain many observations of such galaxies per night, increasing their chances of finding SNe, but it biases them against finding SNe in fainter hosts. Other professional surveys continue to discover SNe with smaller angular separations than amateurs (median value of 11′′.8 vs. 16′′.5), but show a similar median offset in terms of physical separation (5.0 kpc for professionals, 5.2 kpc for amateurs). ASAS-SN continues to be less biased against discoveries close to the host nucleus than either comparison group, as ASAS-SN discoveries show median offsets of 5′′.0 and 2.6 kpc.

These trends are more easily visible when looking at the cumulative distributions of the host galaxy magnitudes and offsets from host nuclei, as shown in Figures 6.6 and 6.7. The distributions clearly show that the ASAS-SN and other professional samples stand out from the amateur sample in host galaxy luminosity, and that SNe discovered by ASAS-SN are more concentrated towards the centers of their hosts than those discovered by either amateurs or other professionals. While the majority of non-ASAS-SN professional discoveries are made by professional searches that do not use difference imaging (e.g., MASTER, Gaia, CRTS) or that systematically avoid central regions of galaxies (e.g., LOSS), which biases these searches against finding

140 sources close to host nuclei, a larger fraction of other professional discoveries were made by surveys that do use difference imaging in 2016 than in previous years due to the start of the ATLAS survey. ASAS-SN continues to find sources with smaller median offsets than its competitors despite this fact, implying that the avoidance of the central regions of galaxies is still fairly common in surveys other than ASAS-SN, regardless of survey strategy and techniques.

The median host magnitudes are MKS ≃ −22.6, MKS ≃ −22.8, and

MKS ≃ −23.8 for ASAS-SN discoveries, other professional discoveries, and amateur discoveries, respectively. There is a clear distinction between professional surveys (including ASAS-SN) and amateurs in terms of host luminosity, and ASAS-SN discoveries have a fainter median than those of other professional surveys.

The impact ASAS-SN has had on the discovery of bright SNe can be most clearly seen in Figure 6.8, which shows the number of bright SNe discovered per month in each month from 2012 through 2016.Milestones in the ASAS-SN timeline, such as the deployment of our southern unit Cassius and software improvements, are shown on the ﬁgure to help visualize the impact of these hardware and software improvements.

In its first year of operation, ASAS-SN had little effect on the number of bright SNe being discovered per month: the average number of bright SNe discovered per month from 2012 January through 2013 May was 13 with a scatter of 4 SNe per month, and from 2013 June through 2014 May the average was 15 with a scatter of 5 SNe per month. However, the addition of our southern unit Cassius and improvements to our pipeline dramatically impacted our detection efficiency and survey cadence, resulting in a significant increase in the number of SNe discovered per month: since ASAS-SN became operational in both hemispheres, the average number of bright SN discoveries has increased to 20 with a scatter of 5 SNe per month. This indicates that ASAS-SN has increased the rate of bright SNe discovered per month since becoming operational in the southern hemisphere, from ∼ 13 ± 2 SNe per month to ∼ 20 ± 2 SNe per month, and has continued to maintain this increased rate for the last 2.5 years—the addition of the 2016 SNe has only decreased the average number of discoveries by 1 SN per month from the previous 2014+2015

141 sample (Holoien et al. 2017b). ASAS-SN is discovering SNe that otherwise would not be found, allowing us to construct a more complete sample of bright, nearby SNe than was previously possible.

ASAS-SN is also less biased in terms of the locations of its discoveries. Figure 6.9 shows a cumulative normalized histogram of bright SNe discovered in 2014 and 2015 with respect to the sine of their declination. The green dashed line represents what would be expected if SNe were discovered at all declinations equally. As can be seen in the ﬁgure, SNe discovered by non-ASAS-SN sources have a clear bias towards the northern hemisphere: the non-ASAS-SN histogram falls signiﬁcantly below the “no bias” expectation except at very low (sin (Dec) ∼< −0.8) and very high (sin (Dec) ∼> 0.6) declinations. This is not unexpected, as many professional searches are based in the northern hemisphere. The ASAS-SN discoveries make the combined sample (the black line) follow the expected distribution very closely. The primary bias remaining in sky coverage is the Galactic plane.

Figure 6.10 shows the redshift distribution of our full sample, divided by type. There is a clear distinction between the three types shown, with the Type Ia distribution peaking between z = 0.03 and z = 0.035, the Type II distribution peaking between z = 0.01 and z = 0.015, and the Type Ib/Ic distribution peaking between z =0.015 and z =0.02. SNe Ia have a more luminous mean peak luminosity than CCSNe, so this distribution is expected for our magnitude-limited sample.

Finally, Figure 6.11 shows a cumulative histogram of SN peak magnitudes with

13.5 < mpeak < 17.0. The Figure shows ASAS-SN discoveries, ASAS-SN discoveries and SNe recovered by ASAS-SN, and all SNe from our sample separately. While amateur observers still account for a large number of the brightest discoveries (those with mpeak ∼< 14.5; Holoien et al. 2017b,c), ASAS-SN has discovered a signiﬁcant fraction of these very bright SNe in 2015 and 2016, accounting for roughly half of such discoveries in our complete sample. ASAS-SN recovers the vast majority of such very bright cases that it does not discover, showing that it is competitive with amateurs who observe the small number of very low-redshift galaxies with high cadence. ASAS-SN discovered or recovered every SN with mpeak < 14.3 in

142 2016 and accounts for a large fraction of the brightest SNe overall. We discuss the non-recovered cases with mpeak < 15 in §6.3.

Figure 6.11 also illustrates an estimate of the completeness of our sample. We fit a broken power-law (shown with a green dashed line in the Figure) to the magnitudes of the observable SNe brighter than mpeak = 17.01, assuming a Euclidean slope below the break magnitude and a variable slope for higher magnitudes. We derived the parameters of the fit using MCMC methods. For only the SNe discovered by ASAS-SN in our complete sample, the number counts are consistent with the Euclidean slope down to m = 16.34 ± 0.07. We find break magnitudes of m = 16.26 ± 0.06 and m = 16.19 ± 0.09 for the sample of SNe discovered and recovered by ASAS-SN and the sample of all bright SNe, respectively.

We find that the integral completenesses of the three samples relative to Euclidean predictions are 0.97 ± 0.02 (0.68 ± 0.03), 0.94 ± 0.02 (0.65 ± 0.03), and 0.93 ± 0.03 (0.71 ± 0.03) at 16.5 (17.0) mag for the ASAS-SN discovered sample, the ASAS-SN discovered + recovered sample, and the total sample, respectively. The differential completenesses relative to Euclidean predictions are 0.71 ± 0.10 (0.22 ± 0.04), 0.62 ± 0.06 (0.22 ± 0.04), and 0.67 ± 0.05 (0.36 ± 0.04) at 16.5 (17.0) mag, respectively. These results imply that roughly 70% of the SNe brighter than mpeak = 17 are being found, and that 20−30% of the mpeak = 17 SNe are being found, relative to the Euclidean expectation extrapolated from brighter SNe. The Euclidean approximation used here does not take into account deviations from Euclidean geometry, the effects of time dilation on SN rates, or K-corrections, and thus likely modestly underestimates the true completeness for faint SNe. These higher order corrections are necessary for a full analysis of nearby SN rates.

6.3. Examination of Missed Cases

While ASAS-SN has been very successful at discovering bright SNe, there were still many bright SNe discovered in 2014, 2015, and 2016 by amateurs and other professional groups that we did not independently recover in our data. Ideally, we should recover all bright SNe in the ASAS-SN survey area, especially those that have

143 very bright (mpeak ≤ 15) peak magnitudes. Since this is not the case, we performed a retrospective study of all 13 SNe discovered by groups other than ASAS-SN in 2014 August and the very bright SNe discovered throughout 2015 that were missed by ASAS-SN, in order to better understand the reasons why we might fail to recover SNe that should be detectable.

Of the 13 SNe discovered by others in 2014 August, five (PSN J01340299- 0104458, SN 2014cc, SN 2014ce, SN 2014 cw, and SN 2014cy) were independently recovered in ASAS-SN data after discovery. For the purpose of this study, we focus on the eight cases where we did not recover the SN in our data. Of these, two (PSN J02451711+4213503 and SN 2014dd) were discovered within 20 degrees of the Galactic plane. Due to the larger number of stars within the Galactic disk, we excluded fields that were within 20 degrees of the Galactic plane from our survey until 2014 December, when we felt confident that our pipeline was running smoothly and would be able to handle the higher false positive rates likely for this region. Thus, these two SNe were outside the survey area in 2014 August, but would have likely been seen had they happened after these fields were added to our search.

Two of the remaining six missed SNe (SN 2014da and OGLE-2014-SN-067) were flagged as transients during our data search, but were rejected as likely false positives. SN 2014da was visible in our data on the day its discovery was announced, but was not flagged as an existing SN. It was not flagged until nearly 50 days after discovery, and was dismissed as the host galaxy had shown previous variability in our data. The host had also been flagged as a possible transient during commissioning observations in 2012. In this case, the previous detection and rejection of the host biased us against discovering the SN. In the case of OGLE-2014-SN-067, shown in Figure 6.12, the transient is very clear and bright in our data and was flagged as a possible transient 12 days before its announcement by OGLE. However, the host galaxy is faint and multiple stars are nearby, so the transient case was closed as a probable variable star. In this case, if we had continued to monitor the source, it would have been detected in multiple epochs prior to the OGLE announcement, and likely would have been flagged as an SN.

144 The remaining four cases (SN 2014cb, SN 2014cd, SN 2014cj, and MASTER OT J162412.26+091303.0, with peak magnitudes of 16.4, 15.8, 17.0, and 16.6, respectively) were not flagged as transients in our data search. Two of these cases, SN 2014cb and SN 2014cj, were not well-observed around the discovery epoch: the previous observation to the discovery was more than two weeks prior in both cases, and the next was at least 4 days after discovery. Both of these SNe were discovered in fields monitored by our southern unit, and the cadence of observations was likely affected by weather and some minor mechanical issues with Cassius during 2014 August. SN 2014cb is only faintly visible in one epoch, and likely would never have been flagged, but SN 2014cj is somewhat clearer, and likely should have been a recovered case. Conversely, the fields for both SN 2014cd and MASTER OT J162412.26+091303.0 were imaged with good cadence (< 3 days) before and after discovery, but again were not flagged. In both cases, the subtracted images from the time were not very clean, and while the SNe were visible in our data, the PSFs changed from epoch-to-epoch and there are poorly subtracted sources nearby. It is likely that our pipeline filtered these sources as unlikely to be real due to the quality of the data, and thus that they were not viewed by a human until this retrospective study. Since the ASAS-SN pipeline flags thousands of possible sources per night, it is necessary to filter out all but the most likely for review by our team members, and unfortunately that means that we will occasionally miss cases such as these.

We expect that the 8 missed SNe from 2014 August are representative of most of the SNe not recovered in our data, and that vast majority of the rest of the missed SNe shown in Figure 6.8 were missed for reasons similar to those described above. The upshot of this study is that many of these missed cases are likely to be at least recovered, if not discovered, by ASAS-SN under normal operations today. During the period from 2014 May 1 to 2014 December 31, the ﬁelds of 18% (11) of the 62 SNe that were not recovered by ASAS-SN were not observed within a week of discovery. Since the Galactic plane is now part of our survey area, we survey the entire visible sky, and we would not miss candidates simply due to their location. Assuming good weather and no mechanical issues, our cadence should be no more than a few days between observations of a ﬁeld, meaning we are unlikely to completely miss cases due to poor cadence. While it is likely impossible to completely eliminate cases

145 like SN 2014cd and MASTER OT J162412.26+091303.0, our pipeline has now been running for over two years and incorporates various improvements, such as the use of machine learning algorithms (Wo´zniak et al. in press) to help us identify the “borderline” cases, allowing us to follow up more candidates such as these.

This has born out in the observations. From 2014 May 1 through 2014 December 31, ASAS-SN recovered 24% (19/80) of the supernovae discovered by other sources. In 2015, this fraction increased dramatically, to 55% (57/104), and it increased further to 64% (62/97) in 2016, indicating that our recovery rate has indeed improved since 2014. Roughly half of the SNe that were not recovered had peak magnitudes of mpeak ≥ 16.5, which makes them more likely to be missed in our data due to factors such as peaking between ASAS-SN observations or occurring during the full moon, when our survey depth is reduced. While it is clear that there is still room for improvement, ASAS-SN is now recovering more than half of the supernovae discovered by other supernova searches, and we expect this fraction to continue to increase in the future.

We also examined the very bright (mpeak ≤ 15.0) SNe from 2015 that were not recovered by ASAS-SN. These are SNe that should be detectable in ASAS-SN data regardless of observing conditions, and are also the most interesting SNe for follow-up study due to their brightness. Understanding the reasons why we missed such SNe is important to ensure that we miss as few of them as possible going forward. Of the SNe discovered by other professional searches and amateurs in 2015, 13 had mpeak ≤ 15.0. Of these 13, only 2 (SN 2015I and MASTER OT J141023.42-431843.7) were not independently recovered in ASAS-SN data. We examine these two cases in detail here.

SN 2015I was discovered in NGC 2357 on 2015 May 02 by amateur astronomers (CBET 4106), and peaked at a magnitude of 14.0. Unfortunately, the field containing this SN was not observable due to Sun constraints from either of our telescope sites for a significant portion of 2015: we were unable to observe the field between 2015 April 22 and 2015 August 29. While the host galaxy had not quite set for the season at the time of discovery, our northern unit Brutus is constrained in how far west it is able to observe due to the presence of the LCO 2-m telescope in the same enclosure.

146 Since the SN was discovered at the beginning of this timeframe, it had faded beyond our detection limit by the time we were able to observe it. While it is unfortunate that we were not able to observe this SN, there is little that could be done to allow us to discover or recover cases like these.

MASTER OT J141023.42-431843.7 was discovered in NGC 5483 on 2015 December 15 by MASTER (Gorbovskoy et al. 2013), and peaked at a magnitude of 14.4. This SN occurred in an observing field that was transferred to one of the new Cassius cameras in the summer of 2015, when we upgraded our Cassius unit to 4 telescopes. Before performing image subtraction in a field, we first construct a reference image by co-adding numerous high-quality exposures. Because of weather and technical issues at the start of operations with four cameras, almost no images were obtained in good conditions prior to the field setting in 2015 September. When the field was again observable in 2016 January, the SN was present, but we did not have enough high-quality images to build a proper reference image. As a result, new images (containing the SN) would have been rapidly incorporated into the reference image, preventing detection. The SN faded below our detection limits on 2016 January 16. Essentially, this was a case of bad timing, as the SN occurred in a field that was not ready for searching at the time it was observable. The field now has a very good reference image built from images spread over the last year, and the SN is trivially found in multiple epochs if we analyze the data from 2016 January using the current reference image.

While we never want to miss any of the brightest SNe, neither of these cases was observable by ASAS-SN, and the number of very bright SNe we have missed has decreased over time: in the latter 8 months of 2014 we missed 4 SNe with mpeak ≤ 15.0, and we missed only 1 such SN in 2016. This indicates that we have improved our eﬃciency with very bright SNe to the point where we are now highly unlikely to miss them in the future if they are observable by ASAS-SN.

147 6.4. Discussion

Our total combined sample of bright SNe discovered between 2014 May 1 and 2016 December 31 includes 668 SNe, 387 discovered by ASAS-SN. The combined sample remains similar to that of an ideal magnitude-limited sample from Li et al. (2011) with a smaller proportion of SNe Ia relative to CCSNe than expected.

ASAS-SN is the only professional survey that provides a complete, rapid- cadence, all-sky survey of the nearby transient universe, and continues to have a major impact on the discovery and follow-up of bright SNe. Even with the advent of recent professional surveys, amateur astronomers, who focus on bright and nearby galaxies for their SN searches, remain the primary competition to ASAS-SN for new discoveries. Our analyses show that ASAS-SN continues to ﬁnd SNe that would not be found otherwise (e.g., Figure 6.8) and that it ﬁnds SNe closer to galactic nuclei and in less luminous hosts than its competitors (Figures 6.4 and 6.5). In 2016 and 2015 ASAS-SN recovered the majority of bright SNe that it did not discover, and

ASAS-SN discovered or recovered all but one of the very bright (mpeak ≤ 15) SNe that were discovered in 2016.

Figure 6.11 shows that the magnitude distribution of SNe discovered between 2014 May 1 and 2016 December 31 is roughly complete to a peak magnitude of mpeak = 16.2 and that it is roughly 70% complete for mpeak ≤ 17.0. While the analyses presented here cannot be used to determine nearby SN rates, since the absolute normalization of the expected number of SNe has not been addressed (by accounting for the exact sky coverage and time windows, for example), these are the precursors to rate calculations which will be presented in future work by the ASAS-SN team.

Such rate calculations could have a signiﬁcant impact on a number of ﬁelds. Within a few hundred Mpc, the measured CCSN rate is about half as big as expected from star formation rates in the same volume (e.g., Horiuchi et al. 2011, 2013), and new measurements of nearby SNe, particularly from galaxy-blind surveys like ASAS-SN, are needed to address this discrepancy. Furthermore, nearby SNe, besides being the easiest to study in the optical, are also the most promising objects

148 for multi-messenger studies, which could include gravitational waves (e.g., Ando et al. 2013; Nakamura et al. 2016), MeV gamma rays from SNe Ia (e.g., Horiuchi & Beacom 2010; Diehl et al. 2014; Churazov et al. 2015) and GeV–TeV gamma rays and neutrinos from rare types of CCSNe (e.g., Ando & Beacom 2005; Murase et al. 2011; Abbasi et al. 2012). Such joint measurements would greatly increase the scientiﬁc reach of ASAS-SN discoveries.

ASAS-SN continues to discover many of the best and brightest transients in the sky, and there are many ways to use our unbiased sample to impact SN research now and in the future. With an expansion to 20 cameras coming in 2017, ASAS-SN’s impact will only grow, as it will ﬁnd the vast majority of bright transients visible in the night sky.

149 ASAS-SN Discoveries (388)

Ia 77% (298)

3% (10) Ib/Ic

21% (80) II

Fig. 6.1.— Breakdown by type of the SNe discovered by ASAS-SN between 2014 May 01 and 2016 December 31, excluding SLSNe and considering SNe IIb in the “Type II” sample. The proportions of each type is very similar to that of an ideal magnitude-limited sample (Li et al. 2011).

150 Non-ASAS-SN Discoveries (279)

54% (151)

11% (30) Ib/Ic

35% (98)

Fig. 6.2.— Breakdown by type of the SNe discovered by non-ASAS-SN sources between 2014 May 01 and 2016 December 31, excluding SLSNe and considering SNe IIb in the “Type II” sample. There is a signiﬁcantly higher fraction of CCSNe, particularly SNe Ib/Ic, than would be expected for a strictly magnitude-limited sample (Li et al. 2011).

151 Full Sample (667)

67% (449)

6% (40) Ib/Ic

27% (178)

Fig. 6.3.— Breakdown by type of the combined bright SN sample, excluding SLSNe and considering SNe IIb in the “Type II” sample. As with the non-ASAS-SN sample shown in Figure 6.2, there is a higher fraction of CCSNe, than would be expected for a magnitude-limited sample (Li et al. 2011).

152 Host Galaxy log L/L⋆ -3 -2 -1 0 1

100

Offset (arcsec) ASAS-SN Other Prof. 0.1 Amateurs -16 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 -27

Host Galaxy MKs

Fig. 6.4.— Offset from the host nucleus in arcseconds compared to the absolute KS- band host magnitude for all SNe in our combined sample discovered between 2014 May 1 and 2016 December 31. The top axis shows log (L/L⋆) values corresponding to the magnitude range shown on the bottom scale assuming M⋆,KS = −24.2 (Kochanek et al. 2001). ASAS-SN SN discoveries are shown as red stars, amateur discoveries are shown as black circles, and discoveries by other professional searches are shown as blue squares. Triangles indicate upper limits on the host galaxy magnitudes for hosts that were not detected in 2MASS or WISE. Points are filled for SNe that were independently recovered by ASAS-SN. Median offsets and host magnitudes for ASAS- SN discoveries, amateur discoveries, and other professional discoveries are indicated using dashed, dotted, and dash-dotted lines, respectively, in colors that match the data points.

153 Host Galaxy log L/L⋆ -3 -2 -1 0 1

1 Offset (kpc) 0.1 ASAS-SN Other Prof. Amateurs -16 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 -27

Host Galaxy MKs

Fig. 6.5.— Offset from the host nucleus in kpc compared to the absolute KS-band host magnitude for all SNe in our combined sample discovered between 2014 May 1 and 2016 December 31. The top axis shows log (L/L⋆) values corresponding to the magnitude range shown on the bottom scale assuming M⋆,KS = −24.2 (Kochanek et al. 2001). ASAS-SN SN discoveries are shown as red stars, amateur discoveries are shown as black circles, and discoveries by other professional searches are shown as blue squares. Triangles indicate upper limits on the host galaxy magnitudes for hosts that were not detected in 2MASS or WISE. Points are filled for SNe that were independently recovered by ASAS-SN. Median offsets and host magnitudes for ASAS- SN discoveries, amateur discoveries, and other professional discoveries are indicated using dashed, dotted, and dash-dotted lines, respectively, in colors that match the data points.

154 1.0 Amateurs 0.8 Other Prof. ASAS-SN 0.6 SN f 0.4

0.2

0.0 −16 −17 −18 −19 −20 −21 −22 −23 −24 −25 −26

Host Galaxy MKS

Fig. 6.6.— Cumulative, normalized distributions of host galaxy absolute magnitudes for the ASAS-SN SN sample (red), the other professional sample (blue), and the amateur sample (black). Amateur discoveries are clearly more biased towards more luminous hosts than professional surveys (including ASAS-SN).

155 1.0

0.8

0.6 SN f 0.4 Amateurs 0.2 Other Prof. ASAS-SN 0.0 0 6 12 18 24 30 36 42 48 54 60 Offset (arcsec)

1.0

0.8

0.6 SN f 0.4 Amateurs 0.2 Other Prof. ASAS-SN 0.0 1 3 5 7 9 11 13 15 17 19 21 23 25 Offset (kpc)

Fig. 6.7.— Cumulative, normalized distributions of offset from host nucleus in arcseconds (upper panel) and offset from host nucleus in kpc (lower panel) for the ASAS-SN SN sample (red), the other professional sample (blue), and the amateur sample (black). ASAS-SN finds SNe at smaller offsets, regardless of whether offset is measured in arcseconds or kpc.

156 eipee n21 a,adAA-Ndsoeisacutfra le April. at 2014 since for operation month account every became discoveries in ASAS-SN ASAS-SN discoveries SN and since discoverie bright May, month SN discover 2014 bright every SNe in of in of hemispheres number number median The median previous 2012. the this through shows 2010 line timelin from ASAS-SN pink month the dashed no in were t The milestones that Signiﬁcant oth 2012 SNe shown. by blue. and from discovered in yellow, month shown in SNe are shown each red, ASAS-SN are in in data discoveries ASAS-SN shown in are SN recovered discoveries bright ASAS-SN of Histogram 2016. 6.8.— Fig.

NSN /Month 10 15 20 25 30 35 0 5 2012 Jan. Mar. May

Jul. Recovered Not Recovered Discoveries ASAS-SN Sep. Nov. Real-time Analysis

2013 Jan. Begins Mar. May Jul. Upgraded Brutus Sep. Nov. asu Deployed Cassius

157 2014 Jan.

Month Mar. May ahn Learning Machine

Jul. Implemented Sep. Nov. 2015 Jan. asu Upgraded Cassius Mar. May Jul. Sep. Nov. 2016 Jan. Mar. a exceeded has s

rsucsand sources er May eoee by recovered t s afo all of half ast Jul. di each in ed li both in al Sep. r also are e Nov. hrough 1 Other ASAS-SN All 0.75

SN 0.5 N

0.25

0 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 sin(Dec)

Fig. 6.9.— Cumulative normalized histogram of SN discoveries from 2014 and 2015 with respect to the sine of their declination. ASAS-SN discoveries are shown in red, non-ASAS-SN discoveries are shown in blue, and the combined sample is shown in black. The green dashed line represents what would be expected if SNe were equally likely to be discovered at all declinations. Non-ASAS-SN discoveries have a clear northern bias and fall well below the expectation except near the poles, while ASAS- SN discoveries correct this trend.

158 80 Type Ia 70 Type II Type Ib/Ic 60

SN 40 N

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 Redshift

Fig. 6.10.— Histograms of SN redshifts from our complete sample with a bin width of z =0.005. Distributions for Type Ia (red line), Type II (blue line), and Type Ib/Ic (green line) SNe are shown separately, with subtypes (such as SN 1991T-like and SN 1991bg-like SNe Ia) included as part of their parent groups. As expected due to their larger intrinsic brightness, SNe Ia are predominantly found at higher redshifts, while less luminous CCSNe are found at comparatively lower redshifts.

159 1000 All ASAS-SN Discovered + Recovered ASAS-SN Discovered

100 SN N

1 13.5 14 14.5 15 15.5 16 16.5 17 Peak Magnitude

Fig. 6.11.— Cumulative histogram of SN peak magnitudes using a 0.1 magnitude bin width. The distributions for only ASAS-SN discoveries (red line), ASAS-SN discoveries and SNe recovered independently by ASAS-SN (blue line), and all SNe in the sample (black line) are shown separately. The green dashed line shows a broken power-law ﬁt that has been normalized to the complete sample with a Euclidean slope below the break magnitude and a variable slope for fainter sources, and the lavender dashed line shows an extrapolation of the Euclidean slope to m = 17. The sample is roughly 70% complete for mpeak < 17.

160 Fig. 6.12.— Archival DSS image (far left), ASAS-SN V -band reference image (center- left), ASAS-SN 2014 August 05 V -band image (center-right), and ASAS-SN 2014 August 05 subtraction image (far right) of the supernova OGLE-2014-SN-067, which was not discovered or recovered independently by ASAS-SN. The red circle is centered on the position of the supernova and the blue circle is centered on the position of the nearest galaxy listed in NED. Both circles have a radius of 7′′.0. While the nearest cataloged galaxy was over an arcminute away from the supernova position, there is a stellar source located ∼0′′.4 away from the supernova, well within a pixel of the supernova in our images, and the host galaxy is faint compared to the star. While we clearly detected the transient source, this was classiﬁed as a probable variable star and not monitored in later epochs of observation for this reason.

161 References

Abbasi, R., Abdou, Y., Abu-Zayyad, T., et al. 2012, A&A, 539, A60

Ahn C. P., Alexandroﬀ R., Allende Prieto C., et al. 2012, ApJS, 203, 21

Alard C. 2000, AAPS, 144, 363

Alam, S., Albareti, F. D., Allende Prieto, C., et al. 2015, ApJS, 219, 12

Albareti, F. D., Allende Prieto, C., et al. 2016, ArXiv e-prints, arXiv:1608.02013

Alexander K. D., Berger E., Guillochon J., Zauderer B. A., Williams P. K. G. 2016, ApJ, 819, 25

Anderson, J. P., Gonz´alez-Gait´an, S., Hamuy, M., et al. 2014, ApJ, 786, 67

Ando, S., & Beacom, J. F. 2005, Physical Review Letters, 95, 061103

Ando, S., Baret, B., Bartos, I., et al. 2013, Reviews of Modern Physics, 85, 1401

Arcavi, I., Gal-Yam, A., Cenko, S. B., et al. 2012, ApJ, 756L, 30

Arcavi I. et al. 2014, ApJ, 793, 38

Arnaud K. A. 1996, in Astronomical Society of the Paciﬁc Conference Series, Vol. 101, Astronomical Data Analysis Software and Systems V, ed. G. H. Jacoby & J. Barnes, 17

Assef R. J., Stern D., Kochanek C. S. et al., 2013, ApJ, 772, 26

Bade N., Komossa S., Dahlem M. 1996, A&A, 309, L35

Baltay, C., Rabinowitz, D., Hadjiyska, E., et al. 2013, PASP, 125, 683

Barron, J. T., Stumm, C., Hogg, D. W., Lang, D., & Roweis, S. 2008, AJ, 135, 414

Bayless A. J., Pritchard T. A., Roming P. W. A., et al. 2013, ApJ, 764L, 13

Becker R. H., White R. L., Helfand D. J. 1995, ApJ, 450, 559

162 Bloom J. S. et al. 2011, Science, 333, 203

Blondin S., Tonry J. L. 2007, ApJ, 666, 1024

Botticella M. T., Trundle C., Pastorello A., et al. 2010, ApJ, 717L, 52

Breeveld A. A., Curran P. A., Hoversten E. A., et al. 2010, MNRAS, 406, 1687

Brimacombe J. et al. 2015, The Astronomer’s Telegram, 7910, 1

Brown J. S., Shappee B. J., W.-S Holoien T., Stanek K. Z., Kochanek C. S., Prieto J. L. 2016, MNRAS, 462, 3993

Brown, J. S., W.-S Holoien, T., Auchettl, K., et al. 2017, MNRAS, 466, 4904

Brown T. M. et al. 2013, PASP, 125, 1031

Bruzual G. & Charlot S. 2003, MNRAS, 344, 1000

Burrows D. N., Hill J. E., Nousek J. A., et al. 2005, SSR, 120, 165

Burrows D. N. et al. 2011, Nature, 476, 421

Cardelli J. A., Clayton G. C., & Mathis J. S. 1988, ApJ, 329L, 33

Cash W. 1979, ApJ, 228, 939

Cenko S. B. et al. 2012a, ApJ, 753, 77

Cenko S. B. et al. 2012b, MNRAS, 420, 2684

Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, ArXiv e-prints, arXiv:1612.05560

Chornock R. et al. 2014, ApJ, 780, 44

Churazov, E., Sunyaev, R., Isern, J., et al. 2015, ApJ, 812, 62

Conley, A., Guy, J., Sullivan, M., et al. 2011, ApJS, 192, 1

Dai L., McKinney J. C., Miller M. C. 2015, ApJ, 812L, 39

Denney K. D., Peterson B. M., Dietrich M., Vestergaard M., Bentz M. C. 2009, ApJ, 692, 246

Diehl, R., Siegert, T., Hillebrandt, W., et al. 2014, Science, 345, 1162

Dong, S., Shappee, B. J., Prieto, J. L., et al. 2016, Science, 351, 257

Donley J. L., Brandt W. N., Eracleous M., Boller T. 2002, AJ, 124, 1308

163 Drake, A. J., Djorgovski, S. G., Mahabal, A., et al. 2009, ApJ, 696, 870

Dressler A. et al. 2011, PASP, 123, 288

Evans C. R., Kochanek C. S. 1989, ApJ, 346, L13

Fabricant D., Cheimets P., Caldwell N., Geary J. 1998, PASP, 110, 79

Faran, T., Poznanski, D., Filippenko, A. V., et al. 2014a, MNRAS, 442, 844

Faran, T., Poznanski, D., Filippenko, A. V., et al. 2014b, MNRAS, 445, 554

Filippenko, A. V. 1997, ARA&A, 35, 309

Frieman, J. A., Bassett, B., Becker, A., et al. 2008, AJ, 135, 338

French K. D., Arcavi I., Zabludoﬀ A. 2016, ApJ, 818, L21

Gal-Yam A., Ofek E. O., Shemmer O. 2002, MNRAS, 332, L73

Gal-Yam, A., Mazzali, P. A., Manulis, I., & Bishop, D. 2013, PASP, 125, 749

Gaskell C. M., Rojas Lobos P. A. 2014, MNRAS, 438, L36

Gezari S. et al. 2006, ApJ, 653, L25

Gezari S. et al. 2008, ApJ, 676, 944

Gezari S. et al. 2009, ApJ, 698, 1367

Gezari S. et al. 2012, Nature, 485, 217

Gezari S., Chornock R., Lawrence A., Rest A., Jones D. O., Berger E., Challis P. M., Narayan G. 2015, ApJ, 815, L5

Godoy-Rivera, D., Stanek, K. Z., Kochanek, C. S., et al. 2017, MNRAS, 466, 1428

Gorbovskoy, E. S., Lipunov, V. M., Kornilov, V. G., et al. 2013, Astronomy Reports, 57, 233

Goto T. et al. 2003, PASJ, 55, 771

Graham A. W., Erwin P., Caon N., et al. 2001, ApJ, 563, L11

Graham M. L., Sand D. J., Valenti S., et al. 2014, ApJ, 787, 163

Graur, O., Rodney, S. A., Maoz, D., et al. 2014, ApJ, 783, 28

Grupe D., Thomas H.-C., Leighly K. M. 1999, A&A, 350, L31

164 Guillochon J., Manukian H., Ramirez-Ruiz E. 2014, ApJ, 783, 23

Guillochon J., Ramirez-Ruiz E. 2015, ApJ, 809, 166

G¨ultekin K., Richstone D. O., Gebhardt K., et al., 2009, ApJ, 698, 198

Guy, J., Sullivan, M., Conley, A., et al. 2010, A&A, 523, A7

Hill J. E., Burrows D. N., Nousek J. A. et al., 2004, in Flanagan K. A., Siegmund O. H. W., eds., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series Vol. 5165, X-Ray and Gamma-Ray Instrumentation for Astronomy XIII, 217

Hamuy, M., Phillips, M. M., Maza, J., et al. 1995, AJ, 109, 1

Harutyunyan, A. H., Pfahler, P., Pastorello, A., et al. 2008, A&A, 488, 383

Hayasaki K., Stone N., & Loeb A. 2013, MNRAS, 434, 909

Hodgkin, S. T., Wyrzykowski, L., Blagorodnova, N., & Koposov, S. 2013, Philosophical Transactions of the Royal Society of London Series A, 371, 20120239

Holoien T. W.-S., Brimacombe J., Shappee B. J., et al. 2013, The Astronomer’s Telegram, 5346, 1

Holoien T. W.-S. et al. 2014, MNRAS, 445, 3263

Holoien T. W.-S. et al. 2016a, MNRAS, 455, 2918

Holoien, T. W.-S., Prieto, J. L., Pejcha, O., et al. 2016b, Acta Astron., 66, 219

Holoien T. W.-S., et al. 2016c, MNRAS, 463, 3813

Holoien, T. W.-S., Stanek, K. Z., Kochanek, C. S., et al. 2017a, MNRAS, 464, 2672

Holoien, T. W.-S., Brown, J. S., Stanek, K. Z., et al. 2017b, MNRAS, 467, 1098

Holoien, T. W.-S., Brown, J. S., Stanek, K. Z., et al. 2017c, ArXiv e-prints, arXiv:1704.02320

Horiuchi, S., & Beacom, J. F. 2010, ApJ, 723, 329

Horiuchi, S., Beacom, J. F., Kochanek, C. S. et al. 2011, ApJ, 738, 154

Horiuchi, S., Beacom, J. F., Bothwell, M. S., & Thompson, T. A. 2013, ApJ, 769, 113

Iben, Jr., I., & Tutukov, A. V. 1984, ApJS, 54, 335

165 Inserra C., Smartt S. J., Jerkstrand A., et al. 2013, ApJ, 770, 128

Ivezic, Z. et al. 2008, ArXiv e-prints, arXiv:0805.2366

Jose J. et al. 2014, The Astronomer’s Telegram, 6777, 1

Jones D. O., Riess A. G., Scolnic D. M. 2015, ApJ, 812, 31

Kaiser, N., Aussel, H., Burke, B. E., et al. 2002, in Proceedings of the SPIE, Vol. 4836, Survey and Other Telescope Technologies and Discoveries, ed. J. A. Tyson & S. Wolﬀ, 154–164

Kalberla P. M. W., Burton W. B., Hartmann D., et al. 2005, ˚a, 400, 775

Kim D.-W., Fabbiano G. 2004, ApJ, 611, 846

Kochanek C. S. 1994, ApJ, 422, 508

Kochanek, C. S., Pahre, M. A., Falco, E. E., et al. 2001, ApJ, 560, 566

Kochanek C. S. 2016a, MNRAS, 458, 127

Kochanek C. S. 2016b, MNRAS, 461, 371

Komossa S., Greiner J. 1999, A&A, 349, L45

Komossa S. 2015, Journal of High Energy Astrophysics, 7, 148

Kriek M., van Dokkum P. G, Labb´eI., et al. 2009, ApJ, 700, 221

Lacy J. H., Townes C. H., Hollenbach D. J. 1982, ApJ, 262, 120

Lang, D., Hogg, D. W., Mierle, K., Blanton, M., & Roweis, S. 2010, AJ, 139, 1782

Lantz B. et al. 2004, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 5249, Optical Design and Engineering, ed. L. Mazuray, P. J. Rogers, & R. Wartmann, 146–155

Law, N. M., Kulkarni, S. R., Dekany, R. G., et al. 2009, PASP, 121, 1395

Leloudas, G., Fraser, M., Stone, N. C., et al. 2016, Nature Astronomy, 1, 0002

Leonard D. C. et al. 2002, AJ, 124, 2490

Li, W. D., Filippenko, A. V., Treﬀers, R. R., et al. 2000, in American Institute of Physics Conference Series, Vol. 522, American Institute of Physics Conference Series, ed. S. S. Holt & W. W. Zhang, 103–106

Li W. et al. 2011, MNRAS, 412, 1441

166 Lodato G., Rossi E. M. 2011, MNRAS, 410, 359

Loeb A., Ulmer A. 1997, ApJ, 489, 573

Lyman J. D., Bersier D., James P. A. 2014, MNRAS, 437, 3848

MacLeod M., Guillochon J., & Ramirez-Ruiz E. 2012, ApJ, 757, 134

Magorrian J., Tremaine S. 1999, MNRAS, 309, 447

Maksym W. P., Ulmer M. P., Eracleous M. 2010, ApJ, 722, 1035

Martini P. et al. 2011, PASP, 123, 187

McConnell N. J. & Ma C.-P. 2013, ApJ, 764, 184

Mendel J. T., Simard L., Palmer M., et al. 2014, ApJS, 210, 3

Metzger B. D., Stone N. C. 2016, MNRAS, 461, 948

Miller A. A., Chornock R., Perley D. A., et al. 2009, ApJ, 690, 1303

Miller J. M. et al. 2015, Nature, 526, 542

Modjaz M. et al. 2014, AJ, 147, 99

Morrell N., Prieto J. L. 2013, The Astronomer’s Telegram, 5353, 1

Morrissey, P., Conrow, T., Barlow, T. A., et al. 2007, ApJS, 173, 682

Murase, K., Thompson, T. A., Lacki, B. C., & Beacom, J. F. 2011, Physical Review D, 84, 043003

Nakamura, K., Horiuchi, S., Tanaka, M., et al. 2016, MNRAS, 461, 3296

National Research Council 2010, New Worlds, New Horizons in Astronomy and Astrophysics (Washington, DC: The National Academies Press), doi:10.17226/12951

Nomoto, K. 1982, ApJ, 253, 798

Osterbrock D. E. 1989, Astrophysics of Gaseous Nebulae and Active Galactic Nuclei

Pasham D. R. et al. 2015, ApJ, 805, 68

Pejcha, O., & Prieto, J. L. 2015a, ApJ, 799, 215

Pejcha, O., & Prieto, J. L. 2015b, ApJ, 806, 225

Perlmutter, S., Aldering, G., Goldhaber, G., et al. 1999, ApJ, 517, 565

167 Peterson B. M. et al. 2004, ApJ, 613, 682

Phillips, M. M. 1993, ApJ, 413, L105

Phinney E. S. 1989, Nature, 340, 595

Pignata, G., Maza, J., Antezana, R., et al. 2009, in American Institute of Physics Conference Series, Vol. 1111, American Institute of Physics Conference Series, ed. G. Giobbi, A. Tornambe, G. Raimondo, M. Limongi, L. A. Antonelli, N. Menci, & E. Brocato, 551–554

Piran T., Svirski G., Krolik J., Cheng R. M., Shiokawa H. 2015, ApJ, 806, 164

Pogge R. W., Atwood B., Brewer D. F., et al., 2010, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, 7735, 77350

Poole T. S., Breeveld, A. A. and Page, M. J., et al. 2008, MNRAS, 383, 627

Poznanski D., Kostrzewa-Rutkowska, Z., Wyrzykowski, L., & Blagorodnova, N. 2008, MNRAS, 449, 1753

Prentice S. et al. 2015, The Astronomer’s Telegram, 7936, 1

Prieto J. L., Bersier D., Holoien T. W.-S., et al. 2014, The Astronomer’s Telegram, 5831, 1

Quimby, R. M. 2006, PhD thesis, The University of Texas at Austin

Quimby R. M., Wheeler J. C., Höflich P., Akerlof C. W., Brown P. J., Rykoff E. S. 2007, ApJ, 666, 1093

Rau A. et al. 2009, PASP, 121, 1334

Rees M. J. 1988, Nature, 333, 523

Riess, A. G., Press, W. H., & Kirshner, R. P. 1995, ApJ, 438, L17

Riess, A. G., Filippenko, A. V., Challis, P., et al. 1998, AJ, 116, 1009

Roming P. W. A. et al. 2005, SSR, 120, 95

Roth N., Kasen D., Guillochon J., Ramirez-Ruiz E. 2016, ApJ, 827, 3

Sako M. et al. 2014, arXiv e-prints, arXiv:1401.3317

Sanders, N. E., Soderberg, A. M., Gezari, S., et al. 2015, ApJ, 799, 208

Saxton R. D., Read A. M., Esquej P., Komossa S., Dougherty S., Rodriguez-Pascual P., Barrado D. 2012, A&A, 541, A106

168 Schlaﬂy E. F. & Finkbeiner D. P. 2011, ApJ, 737, 103

Shappee B. J. et al. 2014, ApJ, 788, 48

Shappee, B. J., Piro, A. L., Holoien, T. W.-S., et al. 2016a, ApJ, 826, 144

Shappee, B. J., Piro, A. L., Stanek, K. Z., et al. 2016b, ArXiv e-prints, arXiv:1610.07601

Shappee, B. J., Stanek, K. Z., Kochanek, C. S., & Garnavich, P. M. 2017, ApJ, 841, 48

Shakura N. I. & Sunyaev R. A. 1973, A&A, 24, 337

Shankar F. 2013, Classical and Quantum Gravity, 30, 244001

Shiokawa H., Krolik J. H., Cheng R. M., Piran T., Noble S. C. 2015, ApJ, 804, 85

Skrutskie M. F., Cutri R. M., Stiening R., et al. 2006, AJ, 131, 1163

Smartt S. J. et al. 2015, A&A, 579, A40

Springob C. M., Haynes M. P., Giovanelli R., Kent B. R. 2005, ApJS, 160, 149

Steele I. A., Smith R. J., Rees P. C., et al. 2004, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series Vol. 5489, Ground-based Telescopes, ed. J. M. Oschmann Jr., 679

Stern D. et al. 2004, ApJ, 612, 690

Stoll R., Prieto J. L., Stanek K. Z., et al. 2011, ApJ, 730, 34

Stone N. C., Metzger B. D. 2016, MNRAS, 455, 859

Strateva I. V., Strauss M. A., Hao L., et al. 2003, AJ, 126, 1720

Strubbe L. E., Quataert E. 2009, MNRAS, 400, 2070

Strubbe L. E., Murray N. 2015, MNRAS, 454, 2321

Svirski G., Piran T., Krolik J. 2017, MNRAS, 467, 1426

Tonry, J. L. 2011, PASP, 123, 58

Tutukov, A. V., & Yungelson, L. R. 1979, Acta Astron., 29, 665

Ulmer A. 1999, ApJ, 514, 180

Valenti S. et al. 2014, MNRAS, 438, L101

169 van Velzen S. et al. 2011, ApJ, 741, 73 van Velzen S., Farrar G. R. 2014, ApJ, 792, 53 van Velzen S. et al. 2016, Science, 351, 62

Vink´oJ. et al. 2015, ApJ, 798, 12

Voges W., et al. 1999, A&A, 349, 389

Wang, B., & Han, Z. 2012, New Astronomy Reviews, 56, 122

Webb N. A., Godet O., Wiersema K., et al. 2014, ApJ, 780L, 9

Webbink, R. F. 1984, ApJ, 277, 355

Whelan, J., & Iben, Jr., I. 1973, ApJ, 186, 1007

Wright E. L., Eisenhardt P. R. M., Mainzer A. K., et al. 2010, AJ, 140, 1868

Wyrzykowski,L., Kostrzewa-Rutkowska, Z., Koz lowski, S., et al. 2014, Acta Astron., 64, 197

Yaron O., Gal-Yam A. 2012, PASP, 124, 668

Yun M. S., Reddy N. A., Condon J. J. 2001, ApJ, 554, 803

Zhang T., Wang X., Wu C., et al. 2012, AJ, 144, 131

Zwicky, F. 1938, ApJ, 88, 529

Zwicky, F. 1942, ApJ, 96, 28

170 Appendix A: Follow-up Photometry

This appendix contains all follow-up photometric observations of ASASSN-14ae, ASASSN-14li, ASASSN-15oi, and ASASSN-13co.

171 Magnitude MJD Magnitudea Uncertainty Filter Telescope

56686.07742 16.147 0.022 z LT 56694.00965 16.145 0.027 z LT 56695.05779 16.149 0.025 z LT 56696.15246 16.141 0.027 z LT 56697.01584 16.130 0.033 z LT 56698.03157 16.120 0.026 z LT 56698.97949 16.115 0.028 z LT 56699.95359 16.146 0.023 z LT 56700.97094 16.135 0.030 z LT 56701.93490 16.146 0.027 z LT 56710.02390 16.190 0.028 z LT 56711.04107 16.210 0.023 z LT 56712.15423 16.254 0.033 z LT 56713.07523 16.228 0.026 z LT 56715.02036 16.264 0.027 z LT 56721.88313 16.325 0.029 z LT 56723.86738 16.326 0.035 z LT 56728.01997 16.337 0.030 z LT 56731.89773 16.353 0.032 z LT 56733.90845 16.379 0.027 z LT 56735.92106 16.335 0.034 z LT 56739.98927 16.381 0.029 z LT 56741.89896 16.428 0.030 z LT 56743.98572 16.402 0.030 z LT 56751.94146 16.428 0.031 z LT 56753.94364 16.431 0.028 z LT 56755.04548 16.419 0.031 z LT 56761.90671 16.450 0.058 z LT 56762.90040 16.439 0.031 z LT 56768.92797 16.438 0.035 z LT 56770.00931 16.432 0.030 z LT 56770.95330 16.445 0.025 z LT

(cont’d) Table A.1. Follow-up Photometric Measurements of ASASSN-14ae.

172 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56771.97739 16.481 0.028 z LT 56684.12208 16.245 0.022 i LT 56685.08659 16.237 0.018 i LT 56686.07649 16.199 0.020 i LT 56689.38424 16.211 0.036 i LCO 56692.20445 16.219 0.097 i LCO 56694.00872 16.202 0.021 i LT 56695.05687 16.195 0.022 i LT 56696.15155 16.189 0.023 i LT 56696.17710 16.217 0.047 i LCO 56697.01491 16.192 0.026 i LT 56698.03063 16.200 0.021 i LT 56698.97856 16.201 0.021 i LT 56699.95267 16.196 0.024 i LT 56700.97002 16.203 0.025 i LT 56701.93398 16.206 0.025 i LT 56710.02297 16.283 0.021 i LT 56711.04013 16.275 0.021 i LT 56712.15330 16.328 0.025 i LT 56713.07431 16.327 0.021 i LT 56715.01944 16.352 0.021 i LT 56721.88220 16.420 0.021 i LT 56723.86644 16.427 0.023 i LT 56728.01904 16.487 0.024 i LT 56731.89681 16.514 0.026 i LT 56733.90752 16.493 0.025 i LT 56735.92014 16.508 0.025 i LT 56739.98834 16.520 0.022 i LT 56741.89803 16.590 0.022 i LT 56743.98480 16.570 0.023 i LT 56751.94053 16.565 0.025 i LT 56753.94272 16.577 0.024 i LT

(cont’d)

173 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56755.04456 16.448 0.024 i LT 56761.90579 16.618 0.026 i LT 56762.89948 16.597 0.024 i LT 56768.92704 16.609 0.023 i LT 56770.00838 16.614 0.026 i LT 56770.95238 16.608 0.021 i LT 56684.12117 16.386 0.020 r LT 56685.08567 16.355 0.021 r LT 56686.07558 16.340 0.019 r LT 56689.38237 16.340 0.030 r LCO 56692.20258 16.374 0.045 r LCO 56694.00780 16.369 0.022 r LT 56695.05595 16.369 0.020 r LT 56696.15063 16.371 0.020 r LT 56696.17523 16.362 0.034 r LCO 56697.01400 16.362 0.023 r LT 56698.02972 16.345 0.021 r LT 56698.97764 16.401 0.022 r LT 56699.95175 16.374 0.021 r LT 56700.96910 16.412 0.027 r LT 56701.93306 16.429 0.022 r LT 56710.02204 16.530 0.021 r LT 56711.03922 16.539 0.020 r LT 56712.15239 16.579 0.018 r LT 56713.07339 16.574 0.020 r LT 56715.01852 16.635 0.020 r LT 56721.88128 16.708 0.021 r LT 56723.86553 16.708 0.021 r LT 56728.01812 16.762 0.024 r LT 56731.89589 16.821 0.027 r LT 56733.90660 16.808 0.024 r LT 56735.91922 16.830 0.022 r LT

(cont’d)

174 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56739.98740 16.865 0.021 r LT 56741.89711 16.876 0.022 r LT 56743.98387 16.839 0.022 r LT 56751.93961 16.879 0.022 r LT 56753.94180 16.883 0.022 r LT 56755.04360 16.907 0.023 r LT 56761.90486 16.886 0.029 r LT 56762.89856 16.936 0.026 r LT 56768.92612 16.930 0.020 r LT 56770.00746 16.947 0.021 r LT 56770.95146 16.943 0.022 r LT 56771.97555 16.968 0.020 r LT 56682.51157 16.300 0.100 V ASAS-SN 56684.90883 16.420 0.041 V Swift 56687.03884 16.330 0.100 V Swift 56697.70773 16.540 0.081 V Swift 56702.77099 16.390 0.110 V Swift 56707.92641 16.530 0.130 V Swift 56729.12832 17.070 0.170 V Swift 56734.38386 16.730 0.170 V Swift 56738.53238 17.230 0.120 V Swift 56739.39868 17.070 0.110 V Swift 56744.18546 17.180 0.170 V Swift 56749.78202 16.940 0.170 V Swift 56755.31624 17.290 0.180 V Swift 56760.19089 17.160 0.150 V Swift 56763.39081 17.000 0.150 V Swift 56770.05206 17.470 0.260 V Swift 56825.86171 17.620 0.240 V Swift 56828.44541 17.080 0.150 V Swift 56684.12022 16.533 0.023 g LT 56684.33405 16.528 0.036 g LCO

(cont’d)

175 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56685.08472 16.516 0.023 g LT 56686.07463 16.509 0.021 g LT 56689.38049 16.503 0.027 g LCO 56692.20070 16.595 0.036 g LCO 56694.00685 16.590 0.022 g LT 56695.05500 16.594 0.022 g LT 56696.14969 16.575 0.023 g LT 56696.17336 16.559 0.041 g LT 56697.01305 16.597 0.025 g LT 56698.02878 16.617 0.020 g LT 56698.97669 16.602 0.022 g LT 56699.95080 16.633 0.029 g LT 56700.96816 16.638 0.022 g LT 56701.93212 16.636 0.041 g LT 56710.02110 16.837 0.022 g LT 56711.03827 16.844 0.022 g LT 56712.15143 16.885 0.023 g LT 56713.07244 16.897 0.021 g LT 56715.01756 16.944 0.022 g LT 56721.88033 17.077 0.022 g LT 56723.86458 17.104 0.024 g LT 56728.01718 17.157 0.025 g LT 56731.89495 17.187 0.037 g LT 56733.90565 17.233 0.029 g LT 56735.91827 17.266 0.023 g LT 56739.98646 17.296 0.023 g LT 56741.89617 17.353 0.024 g LT 56743.98291 17.348 0.025 g LT 56751.93865 17.378 0.024 g LT 56753.94086 17.379 0.024 g LT 56755.04266 17.402 0.025 g LT 56761.90391 17.347 0.035 g LT

(cont’d)

176 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56762.89761 17.436 0.033 g LT 56768.92517 17.452 0.025 g LT 56770.00652 17.445 0.025 g LT 56770.95051 17.442 0.022 g LT 56771.97461 17.468 0.022 g LT 56774.95044 17.466 0.021 g LT 56778.95189 17.470 0.021 g LT 56781.91865 17.481 0.022 g LT 56784.92683 17.487 0.024 g LT 56789.92206 17.498 0.029 g LT 56792.93450 17.495 0.025 g LT 56796.90488 17.522 0.022 g LT 56799.89040 17.500 0.022 g LT 56802.96398 17.525 0.023 g LT 56684.89943 16.690 0.045 B Swift 56687.03569 16.720 0.073 B Swift 56697.70161 16.830 0.054 B Swift 56702.76816 16.800 0.082 B Swift 56707.84216 17.170 0.191 B Swift 56729.12580 17.480 0.112 B Swift 56734.38200 17.710 0.151 B Swift 56738.52596 17.740 0.082 B Swift 56739.39238 17.740 0.082 B Swift 56744.18244 17.650 0.122 B Swift 56749.78476 17.720 0.141 B Swift 56755.31990 17.510 0.112 B Swift 56760.18702 17.900 0.122 B Swift 56763.38771 17.570 0.112 B Swift 56770.05016 17.540 0.141 B Swift 56825.85845 17.930 0.141 B Swift 56828.44126 18.340 0.201 B Swift 56686.07863 16.630 0.026 u LT

(cont’d)

177 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56694.01086 16.817 0.029 u LT 56695.05900 16.848 0.015 u LT 56696.15368 16.820 0.038 u LT 56697.01705 16.873 0.028 u LT 56698.03279 16.896 0.024 u LT 56698.98070 16.848 0.027 u LT 56699.95480 16.965 0.037 u LT 56700.97215 16.928 0.070 u LT 56701.93611 17.001 0.072 u LT 56710.02511 17.259 0.018 u LT 56711.04229 17.240 0.040 u LT 56712.15544 17.290 0.031 u LT 56713.07644 17.302 0.036 u LT 56715.02158 17.411 0.033 u LT 56721.88434 17.740 0.054 u LT 56723.86859 17.658 0.047 u LT 56728.02118 17.817 0.056 u LT 56731.89894 18.025 0.108 u LT 56733.90965 17.881 0.050 u LT 56735.92227 17.992 0.046 u LT 56739.99049 18.123 0.027 u LT 56741.90017 18.242 0.038 u LT 56743.98693 18.260 0.042 u LT 56751.94456 18.362 0.039 u LT 56753.94485 18.337 0.051 u LT 56755.04764 18.372 0.048 u LT 56761.90792 18.471 0.111 u LT 56762.90161 18.625 0.102 u LT 56768.92918 18.604 0.060 u LT 56770.01242 18.584 0.045 u LT 56770.95452 18.623 0.048 u LT 56771.97860 18.564 0.042 u LT

(cont’d)

178 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56774.95245 18.349 0.035 u LT 56774.95492 18.639 0.045 u LT 56778.95389 18.684 0.038 u LT 56778.95636 18.716 0.043 u LT 56781.92065 18.669 0.043 u LT 56781.92313 18.696 0.044 u LT 56784.92883 18.631 0.060 u LT 56784.93130 18.715 0.064 u LT 56789.92406 18.603 0.096 u LT 56792.93650 18.777 0.088 u LT 56792.93898 18.712 0.084 u LT 56796.90687 18.807 0.049 u LT 56796.90934 18.789 0.037 u LT 56799.89240 18.842 0.047 u LT 56799.89487 18.779 0.051 u LT 56802.96599 18.792 0.053 u LT 56802.96846 18.816 0.056 u LT 56684.89753 15.580 0.045 U Swift 56687.03513 15.520 0.063 U Swift 56697.70056 15.770 0.054 U Swift 56702.76765 15.910 0.073 U Swift 56707.84122 16.110 0.063 U Swift 56729.12535 17.000 0.122 U Swift 56734.38165 17.060 0.151 U Swift 56738.52485 17.000 0.082 U Swift 56739.39129 17.090 0.082 U Swift 56744.18191 17.270 0.132 U Swift 56749.78160 17.420 0.161 U Swift 56755.31569 17.390 0.141 U Swift 56760.18634 17.380 0.122 U Swift 56763.38716 17.510 0.141 U Swift 56770.04981 17.890 0.251 U Swift

(cont’d)

179 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56825.85787 18.120 0.211 U Swift 56828.44054 18.230 0.251 U Swift 56684.89377 15.110 0.042 UV W 1 Swift 56687.03357 15.090 0.050 UV W 1 Swift 56697.69751 15.550 0.050 UV W 1 Swift 56702.76625 15.710 0.058 UV W 1 Swift 56707.83852 15.910 0.050 UV W 1 Swift 56729.07071 16.870 0.076 UV W 1 Swift 56734.38074 16.920 0.114 UV W 1 Swift 56738.52164 17.090 0.076 UV W 1 Swift 56739.38815 17.120 0.076 UV W 1 Swift 56744.18042 17.220 0.104 UV W 1 Swift 56749.78435 17.360 0.133 UV W 1 Swift 56755.31935 17.400 0.114 UV W 1 Swift 56760.18441 17.830 0.124 UV W 1 Swift 56763.38562 17.820 0.133 UV W 1 Swift 56768.73185 18.180 0.212 UV W 1 Swift 56770.04886 18.000 0.202 UV W 1 Swift 56794.79258 18.550 0.173 UV W 1 Swift 56799.45558 18.150 0.163 UV W 1 Swift 56804.11631 18.390 0.341 UV W 1 Swift 56809.31304 18.500 0.262 UV W 1 Swift 56814.44711 18.220 0.232 UV W 1 Swift 56825.85625 18.600 0.202 UV W 1 Swift 56828.43848 18.530 0.192 UV W 1 Swift 56684.91259 14.770 0.042 UV M2 Swift 56687.03940 14.780 0.042 UV M2 Swift 56697.70877 15.430 0.042 UV M2 Swift 56702.77148 15.640 0.050 UV M2 Swift 56707.92682 15.850 0.067 UV M2 Swift 56729.12878 16.880 0.076 UV M2 Swift 56734.38421 16.910 0.104 UV M2 Swift

(cont’d)

180 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56738.53348 17.140 0.067 UV M2 Swift 56739.39976 17.000 0.058 UV M2 Swift 56744.18599 17.130 0.085 UV M2 Swift 56749.78046 17.330 0.104 UV M2 Swift 56755.31415 17.600 0.104 UV M2 Swift 56760.19156 17.740 0.095 UV M2 Swift 56763.39135 17.690 0.104 UV M2 Swift 56770.05241 17.710 0.133 UV M2 Swift 56794.59550 > 18.13 N/A UV M2 Swift 56799.37997 18.250 0.182 UV M2 Swift 56804.11453 18.430 0.222 UV M2 Swift 56809.30874 18.600 0.153 UV M2 Swift 56814.44225 18.430 0.133 UV M2 Swift 56825.86227 18.880 0.182 UV M2 Swift 56828.44612 18.570 0.133 UV M2 Swift 56684.90137 15.060 0.042 UV W 2 Swift 56687.03627 15.170 0.050 UV W 2 Swift 56697.70267 15.800 0.042 UV W 2 Swift 56702.76867 15.980 0.058 UV W 2 Swift 56707.92452 16.030 0.058 UV W 2 Swift 56729.12627 16.910 0.085 UV W 2 Swift 56734.38236 17.120 0.104 UV W 2 Swift 56734.38236 17.120 0.104 UV W 2 Swift 56738.52707 17.130 0.058 UV W 2 Swift 56739.39347 17.170 0.067 UV W 2 Swift 56744.18299 17.230 0.085 UV W 2 Swift 56749.78247 17.270 0.104 UV W 2 Swift 56755.31680 17.460 0.095 UV W 2 Swift 56760.18770 17.690 0.095 UV W 2 Swift 56763.38825 17.690 0.104 UV W 2 Swift 56770.05052 17.580 0.124 UV W 2 Swift 56794.59036 18.470 0.104 UV W 2 Swift

(cont’d)

181 Table A.1—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56799.37892 18.150 0.202 UV W 2 Swift 56804.11361 18.490 0.252 UV W 2 Swift 56809.30656 18.230 0.143 UV W 2 Swift 56814.43980 18.770 0.192 UV W 2 Swift 56825.85903 18.930 0.182 UV W 2 Swift 56828.44199 18.650 0.143 UV W 2 Swift

aMagnitudes are presented in the natural system for each ﬁlter: ugriz magnitudes are in the AB system, while Swift ﬁlter magnitudes are in the Vega system.

182 Flux Count MJD Fluxa Uncertaintya Count Rateb Uncertaintyb

56991.5 1.81 0.10 0.35 0.01 56993.5 1.89 0.10 0.33 0.01 56995.5 2.09 0.09 0.33 0.01 56998.5 1.92 0.10 0.41 0.02 57001.5 2.01 0.09 0.39 0.02 57004.5 2.17 0.18 0.41 0.02 57007.5 2.60 0.09 0.45 0.02 57010.5 2.60 0.11 0.48 0.02 57013.5 2.27 0.09 0.43 0.02 57016.5 1.98 0.09 0.45 0.02 57019.5 2.17 0.07 0.40 0.01 57022.5 2.36 0.11 0.33 0.01 57029.5 2.22 0.07 0.40 0.01 57033.5 2.29 0.11 0.36 0.02 57037.5 2.50 0.11 0.39 0.02 57040.5 2.01 0.12 0.35 0.01 57043.5 1.89 0.09 0.36 0.02 57046.5 1.99 0.11 0.36 0.02 57048.5 1.91 0.09 0.33 0.01 57051.5 2.00 0.09 0.30 0.02 57053.5 1.77 0.09 0.31 0.01 57057.5 1.80 0.09 0.28 0.01 57059.5 1.65 0.08 0.29 0.01 57065.5 1.66 0.10 0.20 0.01 57068.5 1.60 0.12 0.24 0.01 57071.5 1.64 0.10 0.25 0.01 57074.5 1.68 0.10 0.24 0.01 57077.5 1.56 0.15 0.19 0.01 57080.5 1.85 0.10 0.24 0.01 57086.9 1.81 0.10 0.27 0.01 57089.3 1.58 0.08 0.30 0.01 57099.4 2.11 0.17 0.25 0.01

(cont’d) Table A.2. Swift XRT Photometry of ASASSN-14li.

183 Table A.2—Continued

Flux Count MJD Fluxa Uncertaintya Count Rateb Uncertaintyb

57102.6 1.71 0.10 0.23 0.01 57109.2 1.77 0.12 0.23 0.01 57111.9 2.34 0.11 0.23 0.01 57114.1 1.48 0.10 0.20 0.01 57117.7 2.03 0.13 0.20 0.01 57120.3 1.51 0.08 0.19 0.01 57123.6 1.70 0.13 0.19 0.01 57126.2 1.53 0.10 0.17 0.01 57129.4 1.62 0.13 0.18 0.01 57132.6 1.08 0.09 0.16 0.01 57136.6 1.05 0.07 0.15 0.01 57139.3 1.38 0.13 0.17 0.01 57147.6 1.05 0.08 0.14 0.01 57150.3 1.72 0.14 0.15 0.01 57153.5 1.39 0.10 0.15 0.01 57156.7 1.18 0.11 0.15 0.01

aAll X-ray ﬂuxes and ﬂux uncertainties are given in units of 10−11 ergs s−1 cm−2. bAll count rates and count rate uncertainties are given in counts s−1. The Swift XRT energy range is 0.3 − 10 keV. Data are not corrected for Galactic extinction.

184 Magnitude MJD Magnitudea Uncertainty Filter Telescope

57007.258 15.013 0.013 z LT 57008.280 15.013 0.014 z LT 57009.275 15.102 0.013 z LT 57011.227 15.087 0.014 z LT 57012.211 15.028 0.014 z LT 57014.176 15.002 0.018 z LT 57016.213 15.027 0.013 z LT 57017.205 15.120 0.014 z LT 57018.164 15.110 0.016 z LT 57019.156 15.095 0.015 z LT 57020.153 15.038 0.022 z LT 57021.156 15.076 0.014 z LT 57023.131 15.035 0.014 z LT 57024.214 15.119 0.015 z LT 57025.277 15.114 0.011 z LT 57026.122 15.033 0.018 z LT 57027.120 15.042 0.017 z LT 57028.200 15.110 0.013 z LT 57029.119 15.043 0.016 z LT 57029.145 15.023 0.016 z LT 57037.238 15.113 0.008 z LT 57038.212 15.115 0.007 z LT 57039.157 15.043 0.007 z LT 57041.138 15.056 0.008 z LT 57042.143 15.059 0.008 z LT 57044.153 15.131 0.006 z LT 57046.083 15.149 0.007 z LT 57048.079 15.159 0.007 z LT 57051.083 15.161 0.017 z LT 57053.056 15.142 0.008 z LT 57055.048 15.129 0.009 z LT 57057.082 15.120 0.009 z LT

(cont’d) Table A.3. Follow-up Photometric Measurements of ASASSN-14li.

185 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57059.043 15.157 0.010 z LT 57065.160 15.135 0.007 z LT 57068.143 15.136 0.007 z LT 56989.467 15.291 0.053 i LCO 57003.460 15.213 0.016 i LCO 57003.462 15.178 0.015 i LCO 57004.468 15.223 0.019 i LCO 57004.471 15.253 0.019 i LCO 57006.414 15.214 0.015 i LCO 57006.417 15.203 0.016 i LCO 57006.438 15.260 0.038 i LCO 57006.441 15.234 0.033 i LCO 57006.471 15.260 0.013 i LCO 57006.473 15.279 0.013 i LCO 57008.280 15.265 0.010 i LT 57009.274 15.259 0.009 i LT 57011.227 15.250 0.009 i LT 57014.082 15.255 0.014 i LCO 57014.089 15.233 0.015 i LCO 57014.093 15.252 0.015 i LCO 57016.076 15.298 0.012 i LCO 57016.079 15.271 0.020 i LCO 57017.205 15.277 0.009 i LT 57017.529 15.284 0.012 i LCO 57017.532 15.296 0.012 i LCO 57018.163 15.265 0.010 i LT 57019.155 15.270 0.009 i LT 57020.319 15.284 0.013 i LCO 57020.322 15.273 0.012 i LCO 57021.155 15.274 0.009 i LT 57021.701 15.316 0.012 i LCO 57021.703 15.287 0.012 i LCO

(cont’d)

186 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57024.214 15.302 0.010 i LT 57025.277 15.286 0.008 i LT 57025.692 15.287 0.018 i LCO 57026.122 15.252 0.017 i LT 57028.042 15.310 0.019 i LCO 57029.145 15.286 0.013 i LT 57029.679 15.296 0.002 i LCO 57029.682 15.299 0.020 i LCO 57032.031 15.288 0.024 i LCO 57037.237 15.321 0.005 i LT 57038.211 15.307 0.005 i LT 57041.007 15.316 0.016 i LCO 57041.010 15.319 0.016 i LCO 57043.003 15.343 0.016 i LCO 57043.006 15.329 0.016 i LCO 57044.152 15.314 0.005 i LT 57045.111 15.298 0.012 i LCO 57045.113 15.321 0.012 i LCO 57046.082 15.318 0.005 i LT 57047.048 15.342 0.013 i LCO 57047.050 15.303 0.013 i LCO 57048.078 15.334 0.005 i LT 57050.994 15.320 0.016 i LCO 57051.082 15.356 0.016 i LT 57053.054 15.321 0.006 i LT 57055.047 15.311 0.007 i LT 57057.081 15.299 0.008 i LT 57059.042 15.276 0.009 i LT 57065.159 15.306 0.005 i LT 57068.142 15.316 0.005 i LT 56989.465 15.404 0.036 r LT 57007.257 15.375 0.008 r LT

(cont’d)

187 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57008.279 15.447 0.008 r LT 57009.274 15.453 0.007 r LT 57011.226 15.459 0.007 r LT 57012.210 15.396 0.007 r LT 57014.174 15.402 0.009 r LT 57016.212 15.415 0.007 r LT 57017.204 15.484 0.008 r LT 57018.163 15.480 0.008 r LT 57019.155 15.461 0.008 r LT 57020.152 15.451 0.012 r LT 57021.154 15.485 0.008 r LT 57024.213 15.506 0.009 r LT 57025.276 15.503 0.008 r LT 57026.121 15.503 0.019 r LT 57028.199 15.511 0.012 r LT 57029.144 15.462 0.012 r LT 57037.236 15.518 0.004 r LT 57038.210 15.526 0.004 r LT 57039.155 15.472 0.004 r LT 57041.136 15.497 0.005 r LT 57042.141 15.493 0.004 r LT 57044.151 15.535 0.004 r LT 57046.081 15.539 0.004 r LT 57048.077 15.546 0.004 r LT 57051.081 15.518 0.015 r LT 57053.053 15.537 0.007 r LT 57055.046 15.551 0.008 r LT 57057.080 15.558 0.009 r LT 57059.041 15.550 0.011 r LT 57065.158 15.540 0.004 r LT 57068.141 15.548 0.004 r LT 56991.431 15.570 0.061 V Swift

(cont’d)

188 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56993.888 15.670 0.061 V Swift 56995.292 15.570 0.061 V Swift 56998.274 15.620 0.061 V Swift 57001.608 15.470 0.051 V Swift 57004.264 15.700 0.100 V Swift 57007.266 15.650 0.061 V Swift 57010.803 15.630 0.061 V Swift 57013.067 15.710 0.061 V Swift 57016.062 15.690 0.061 V Swift 57019.515 15.720 0.071 V Swift 57022.718 15.690 0.061 V Swift 57029.319 15.690 0.081 V Swift 57033.113 15.780 0.081 V Swift 57036.108 15.890 0.081 V Swift 57039.036 15.870 0.071 V Swift 57042.230 15.940 0.091 V Swift 57045.560 15.790 0.071 V Swift 57048.756 15.770 0.081 V Swift 57051.468 15.750 0.081 V Swift 57054.007 15.750 0.081 V Swift 57057.530 15.770 0.061 V Swift 57059.798 15.830 0.091 V Swift 57065.852 15.740 0.071 V Swift 57068.781 15.810 0.061 V Swift 57071.710 15.750 0.061 V Swift 57074.848 15.800 0.081 V Swift 57077.576 15.850 0.071 V Swift 57080.891 15.700 0.110 V Swift 57086.887 15.850 0.071 V Swift 57089.350 15.760 0.061 V Swift 57132.462 15.710 0.081 V Swift 56989.463 15.752 0.041 g LCO

(cont’d)

189 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57003.457 15.715 0.014 g LCO 57004.465 15.740 0.016 g LCO 57006.412 15.782 0.016 g LCO 57006.436 15.797 0.056 g LCO 57006.468 15.794 0.011 g LCO 57007.256 15.798 0.010 g LT 57008.279 15.830 0.009 g LT 57009.273 15.845 0.008 g LT 57011.226 15.840 0.008 g LT 57012.209 15.856 0.009 g LT 57014.174 15.856 0.011 g LT 57016.073 15.887 0.010 g LCO 57016.211 15.857 0.008 g LT 57017.203 15.887 0.008 g LT 57017.526 15.910 0.008 g LCO 57018.162 15.867 0.009 g LT 57019.154 15.873 0.008 g LT 57020.151 15.916 0.015 g LT 57020.317 15.905 0.010 g LCO 57021.154 15.900 0.008 g LT 57021.698 15.931 0.011 g LCO 57023.130 15.946 0.017 g LT 57024.213 15.970 0.012 g LT 57025.275 15.930 0.008 g LT 57025.687 15.900 0.034 g LCO 57026.120 15.889 0.032 g LT 57027.118 15.942 0.024 g LT 57028.040 15.956 0.026 g LCO 57028.198 15.960 0.016 g LT 57029.117 15.948 0.021 g LT 57029.144 15.956 0.018 g LT 57029.677 15.958 0.030 g LCO

(cont’d)

190 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57037.235 15.963 0.004 g LT 57038.209 15.966 0.004 g LT 57039.154 15.968 0.004 g LT 57041.004 15.996 0.014 g LCO 57041.135 15.985 0.004 g LT 57042.140 15.975 0.004 g LT 57043.000 15.977 0.014 g LCO 57044.150 15.990 0.004 g LT 57045.108 15.986 0.010 g LCO 57046.080 15.993 0.004 g LT 57047.042 15.985 0.016 g LCO 57048.076 16.005 0.004 g LT 57051.080 15.974 0.020 g LT 57053.052 15.967 0.011 g LT 57055.045 16.036 0.013 g LT 57057.079 16.001 0.014 g LT 57059.040 16.073 0.017 g LT 57065.157 15.985 0.005 g LT 57068.130 16.006 0.003 g LT 57068.140 16.006 0.004 g LT 57070.125 16.021 0.003 g LT 57077.222 16.069 0.005 g LT 57079.071 16.062 0.003 g LT 57085.032 16.055 0.008 g LT 57093.076 16.082 0.004 g LT 57095.071 16.081 0.003 g LT 57097.994 16.095 0.003 g LT 57109.927 16.116 0.005 g LT 57112.017 16.112 0.009 g LT 57113.968 16.102 0.011 g LT 57117.977 16.110 0.009 g LT 57121.021 16.099 0.004 g LT

(cont’d)

191 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57122.959 16.115 0.003 g LT 57124.975 16.114 0.003 g LT 57127.004 16.114 0.003 g LT 57129.961 16.106 0.003 g LT 57132.982 16.128 0.003 g LT 57135.962 16.125 0.003 g LT 57138.944 16.111 0.004 g LT 57142.113 16.152 0.008 g LT 57147.941 16.117 0.005 g LT 57150.951 16.127 0.003 g LT 57153.912 16.127 0.004 g LT 57158.925 16.105 0.004 g LT 57161.893 16.110 0.003 g LT 56991.426 15.960 0.045 B Swift 56993.882 16.020 0.045 B Swift 56995.275 15.930 0.045 B Swift 56998.228 15.940 0.054 B Swift 57001.602 16.010 0.045 B Swift 57004.132 15.970 0.045 B Swift 57007.260 16.050 0.045 B Swift 57010.798 16.040 0.045 B Swift 57013.061 16.060 0.045 B Swift 57016.056 16.170 0.045 B Swift 57019.510 16.160 0.054 B Swift 57022.712 16.090 0.045 B Swift 57029.315 16.180 0.063 B Swift 57033.110 16.160 0.063 B Swift 57036.045 16.310 0.063 B Swift 57039.032 16.290 0.054 B Swift 57042.228 16.180 0.063 B Swift 57045.555 16.220 0.054 B Swift 57048.752 16.300 0.054 B Swift

(cont’d)

192 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57051.465 16.360 0.063 B Swift 57054.004 16.260 0.063 B Swift 57057.524 16.220 0.045 B Swift 57059.796 16.310 0.063 B Swift 57065.779 16.270 0.063 B Swift 57068.775 16.270 0.045 B Swift 57071.704 16.340 0.045 B Swift 57074.844 16.270 0.054 B Swift 57077.572 16.330 0.054 B Swift 57080.889 16.340 0.082 B Swift 57086.882 16.340 0.054 B Swift 57089.344 16.340 0.045 B Swift 57132.459 16.390 0.063 B Swift 57007.258 16.586 0.059 u LT 57008.281 16.537 0.069 u LT 57009.276 16.585 0.062 u LT 57011.228 16.690 0.059 u LT 57012.212 16.724 0.062 u LT 57014.176 16.761 0.100 u LT 57016.214 16.749 0.060 u LT 57017.206 16.749 0.061 u LT 57018.164 16.784 0.077 u LT 57020.153 16.716 0.134 u LT 57021.156 16.744 0.090 u LT 57024.215 16.842 0.069 u LT 57025.278 16.842 0.052 u LT 57029.146 16.844 0.125 u LT 57037.239 16.976 0.033 u LT 57038.213 17.054 0.030 u LT 57039.158 17.050 0.033 u LT 57041.139 16.970 0.044 u LT 57042.144 17.070 0.037 u LT

(cont’d)

193 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57044.154 17.036 0.030 u LT 57046.084 17.024 0.049 u LT 57048.080 17.034 0.045 u LT 57051.084 16.995 0.117 u LT 57053.057 17.106 0.077 u LT 57057.083 17.069 0.120 u LT 57059.044 16.981 0.223 u LT 57065.161 17.090 0.034 u LT 57068.132 17.126 0.014 u LT 57068.144 17.130 0.030 u LT 57070.128 17.126 0.015 u LT 57077.224 17.290 0.033 u LT 57079.073 17.258 0.017 u LT 57085.034 17.332 0.044 u LT 57093.078 17.370 0.019 u LT 57095.073 17.494 0.014 u LT 57097.996 17.518 0.015 u LT 57109.929 17.478 0.039 u LT 57112.019 17.391 0.045 u LT 57117.979 17.458 0.044 u LT 57121.023 17.526 0.022 u LT 57122.961 17.564 0.019 u LT 57124.977 17.510 0.017 u LT 57127.006 17.538 0.016 u LT 57129.963 17.540 0.017 u LT 57132.984 17.543 0.017 u LT 57135.965 17.524 0.017 u LT 57138.946 17.507 0.022 u LT 57142.115 17.484 0.040 u LT 57144.907 17.565 0.055 u LT 57150.953 17.543 0.017 u LT 57158.927 17.562 0.021 u LT

(cont’d)

194 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57161.895 17.572 0.018 u LT 56991.425 15.150 0.045 U Swift 56993.881 15.260 0.045 U Swift 56995.274 15.220 0.045 U Swift 56998.227 15.260 0.045 U Swift 57001.601 15.280 0.045 U Swift 57004.131 15.400 0.045 U Swift 57007.259 15.470 0.045 U Swift 57010.797 15.490 0.045 U Swift 57013.060 15.420 0.045 U Swift 57016.055 15.530 0.045 U Swift 57019.510 15.600 0.054 U Swift 57022.711 15.530 0.045 U Swift 57029.315 15.750 0.063 U Swift 57033.109 15.800 0.063 U Swift 57036.044 15.780 0.063 U Swift 57039.031 15.860 0.063 U Swift 57042.227 15.800 0.073 U Swift 57045.555 15.880 0.063 U Swift 57048.751 15.860 0.063 U Swift 57051.464 15.870 0.063 U Swift 57054.003 15.950 0.073 U Swift 57057.523 15.940 0.054 U Swift 57059.795 15.880 0.073 U Swift 57065.778 15.960 0.073 U Swift 57068.774 15.990 0.054 U Swift 57071.703 15.950 0.054 U Swift 57074.844 15.990 0.063 U Swift 57077.571 16.030 0.063 U Swift 57080.889 16.050 0.102 U Swift 57086.881 16.170 0.063 U Swift 57089.343 16.230 0.063 U Swift

(cont’d)

195 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57099.394 16.190 0.054 U Swift 57102.391 16.170 0.082 U Swift 57108.771 16.270 0.073 U Swift 57111.902 16.280 0.054 U Swift 57114.093 16.330 0.073 U Swift 57117.692 16.350 0.054 U Swift 57120.285 16.340 0.054 U Swift 57123.547 16.380 0.054 U Swift 57126.138 16.430 0.073 U Swift 57129.202 16.400 0.073 U Swift 57132.458 16.530 0.092 U Swift 57136.119 16.470 0.082 U Swift 57139.317 16.450 0.054 U Swift 57147.565 16.480 0.054 U Swift 57150.228 16.340 0.045 U Swift 57153.420 16.450 0.063 U Swift 57156.351 16.540 0.102 U Swift 56991.423 14.730 0.042 UV W 1 Swift 56993.878 14.980 0.042 UV W 1 Swift 56995.271 14.820 0.042 UV W 1 Swift 56998.225 14.880 0.050 UV W 1 Swift 57001.598 14.890 0.042 UV W 1 Swift 57004.128 14.940 0.042 UV W 1 Swift 57007.256 15.150 0.042 UV W 1 Swift 57010.794 15.240 0.042 UV W 1 Swift 57013.057 15.210 0.042 UV W 1 Swift 57016.052 15.340 0.050 UV W 1 Swift 57019.507 15.400 0.050 UV W 1 Swift 57022.708 15.390 0.050 UV W 1 Swift 57029.313 15.620 0.058 UV W 1 Swift 57033.038 15.690 0.058 UV W 1 Swift 57036.042 15.640 0.050 UV W 1 Swift

(cont’d)

196 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57039.029 15.730 0.050 UV W 1 Swift 57042.226 15.750 0.067 UV W 1 Swift 57045.552 15.820 0.058 UV W 1 Swift 57051.463 15.780 0.058 UV W 1 Swift 57054.002 15.780 0.067 UV W 1 Swift 57057.520 15.770 0.050 UV W 1 Swift 57059.794 15.880 0.067 UV W 1 Swift 57065.776 15.920 0.058 UV W 1 Swift 57068.771 16.080 0.058 UV W 1 Swift 57071.700 15.930 0.050 UV W 1 Swift 57074.842 16.060 0.067 UV W 1 Swift 57077.569 16.160 0.067 UV W 1 Swift 57080.888 16.110 0.085 UV W 1 Swift 57086.878 16.250 0.058 UV W 1 Swift 57089.340 16.340 0.058 UV W 1 Swift 57099.391 16.450 0.067 UV W 1 Swift 57102.390 16.430 0.095 UV W 1 Swift 57105.254 16.700 0.143 UV W 1 Swift 57108.770 16.520 0.085 UV W 1 Swift 57111.899 16.500 0.058 UV W 1 Swift 57114.092 16.440 0.085 UV W 1 Swift 57117.688 16.470 0.058 UV W 1 Swift 57120.282 16.620 0.067 UV W 1 Swift 57123.544 16.640 0.067 UV W 1 Swift 57126.136 16.740 0.085 UV W 1 Swift 57129.201 16.880 0.085 UV W 1 Swift 57132.457 16.650 0.085 UV W 1 Swift 57136.118 16.680 0.095 UV W 1 Swift 57139.314 16.740 0.067 UV W 1 Swift 57147.562 16.750 0.067 UV W 1 Swift 57150.224 16.770 0.058 UV W 1 Swift 57153.418 16.730 0.067 UV W 1 Swift

(cont’d)

197 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57156.350 16.740 0.114 UV W 1 Swift 56991.432 14.460 0.042 UV M2 Swift 56993.889 14.910 0.042 UV M2 Swift 56995.283 14.540 0.042 UV M2 Swift 56998.275 14.680 0.036 UV M2 Swift 57001.609 14.680 0.042 UV M2 Swift 57004.264 14.730 0.050 UV M2 Swift 57007.268 14.900 0.042 UV M2 Swift 57010.804 14.940 0.042 UV M2 Swift 57013.068 14.950 0.042 UV M2 Swift 57016.063 15.140 0.042 UV M2 Swift 57019.516 15.170 0.042 UV M2 Swift 57022.719 15.180 0.042 UV M2 Swift 57029.319 15.580 0.050 UV M2 Swift 57033.114 15.370 0.050 UV M2 Swift 57036.109 15.490 0.050 UV M2 Swift 57039.037 15.500 0.050 UV M2 Swift 57042.231 15.620 0.058 UV M2 Swift 57045.561 15.620 0.050 UV M2 Swift 57051.469 15.660 0.050 UV M2 Swift 57054.007 15.670 0.050 UV M2 Swift 57057.531 15.740 0.050 UV M2 Swift 57059.799 15.780 0.058 UV M2 Swift 57065.853 15.700 0.076 UV M2 Swift 57068.782 16.070 0.050 UV M2 Swift 57071.711 15.870 0.050 UV M2 Swift 57074.848 15.950 0.058 UV M2 Swift 57077.577 16.020 0.050 UV M2 Swift 57080.891 16.120 0.076 UV M2 Swift 57086.888 16.130 0.050 UV M2 Swift 57089.351 16.300 0.050 UV M2 Swift 57099.388 16.310 0.067 UV M2 Swift

(cont’d)

198 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57102.389 16.600 0.104 UV M2 Swift 57105.250 16.450 0.058 UV M2 Swift 57108.768 16.490 0.085 UV M2 Swift 57111.895 16.580 0.067 UV M2 Swift 57114.090 16.570 0.095 UV M2 Swift 57117.683 16.560 0.058 UV M2 Swift 57120.278 16.700 0.067 UV M2 Swift 57123.540 16.670 0.067 UV M2 Swift 57126.134 16.770 0.085 UV M2 Swift 57129.198 16.810 0.085 UV M2 Swift 57132.462 16.740 0.076 UV M2 Swift 57136.116 16.790 0.104 UV M2 Swift 57139.309 16.670 0.067 UV M2 Swift 57147.558 16.790 0.067 UV M2 Swift 57150.218 16.850 0.067 UV M2 Swift 57153.414 16.750 0.076 UV M2 Swift 57156.349 16.860 0.124 UV M2 Swift 56991.427 14.260 0.042 UV W 2 Swift 56993.883 14.700 0.042 UV W 2 Swift 56995.276 14.310 0.042 UV W 2 Swift 56998.229 14.360 0.050 UV W 2 Swift 57001.603 14.460 0.042 UV W 2 Swift 57004.133 14.480 0.042 UV W 2 Swift 57007.261 14.690 0.042 UV W 2 Swift 57010.799 14.770 0.042 UV W 2 Swift 57013.062 14.730 0.042 UV W 2 Swift 57016.057 14.920 0.042 UV W 2 Swift 57019.511 15.010 0.042 UV W 2 Swift 57022.713 14.990 0.042 UV W 2 Swift 57029.316 15.380 0.050 UV W 2 Swift 57033.111 15.170 0.050 UV W 2 Swift 57036.046 15.560 0.058 UV W 2 Swift

(cont’d)

199 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57039.033 15.350 0.042 UV W 2 Swift 57042.228 15.400 0.050 UV W 2 Swift 57045.556 15.510 0.050 UV W 2 Swift 57051.466 15.440 0.050 UV W 2 Swift 57054.004 15.550 0.050 UV W 2 Swift 57057.525 15.570 0.042 UV W 2 Swift 57059.796 15.580 0.050 UV W 2 Swift 57065.779 15.550 0.085 UV W 2 Swift 57068.776 16.050 0.050 UV W 2 Swift 57071.705 15.740 0.042 UV W 2 Swift 57074.845 15.770 0.050 UV W 2 Swift 57077.573 15.800 0.050 UV W 2 Swift 57080.890 15.940 0.067 UV W 2 Swift 57086.883 16.030 0.050 UV W 2 Swift 57089.345 16.120 0.050 UV W 2 Swift 57099.395 16.250 0.050 UV W 2 Swift 57102.392 16.360 0.067 UV W 2 Swift 57108.772 16.310 0.067 UV W 2 Swift 57111.904 16.450 0.050 UV W 2 Swift 57114.094 16.350 0.067 UV W 2 Swift 57117.694 16.480 0.058 UV W 2 Swift 57120.286 16.690 0.058 UV W 2 Swift 57123.549 16.630 0.058 UV W 2 Swift 57126.139 16.760 0.067 UV W 2 Swift 57129.203 16.830 0.076 UV W 2 Swift 57132.459 16.660 0.067 UV W 2 Swift 57136.120 16.660 0.076 UV W 2 Swift 57139.318 16.590 0.050 UV W 2 Swift 57147.567 16.670 0.050 UV W 2 Swift 57150.230 16.720 0.050 UV W 2 Swift

(cont’d)

200 Table A.3—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57153.422 16.790 0.058 UV W 2 Swift 57156.352 16.680 0.085 UV W 2 Swift

aMagnitudes are presented in the natural system for each ﬁlter: ugriz magnitudes are in the AB system, while Swift ﬁlter magnitudes are in the Vega system.

Lower Upper Epochs MJD Fluxa Uncertaintya Uncertaintya Combinedb

57263.0 4.18 2.04 1.73 01 − 04 57279.0 6.80 2.04 1.73 05 − 09 57296.0 6.77 2.31 1.97 10 − 13 57306.0 8.32 2.98 2.35 14 − 16 57319.0 6.80 2.25 1.88 17 − 20 57335.0 7.79 2.20 1.88 21 − 24 57345.0 6.17 2.98 2.35 25 − 27

aAll X-ray ﬂuxes and ﬂux uncertainties are given in units of 10−14 ergs s−1 cm−2. bThe “Epochs Combined” column indicates which Swift epochs were combined to obtain the listed measurement. The Swift XRT energy range is 0.3 − 10 keV. Data are not corrected for Galactic extinction.

Table A.4. Swift XRT Photometry of ASASSN-15oi.

201 Magnitude MJD Magnitudea Uncertainty Filter Telescope

57249.532 15.45 0.129 I LCO 57252.002 15.54 0.158 I LCO 57256.428 15.46 0.133 I LCO 57258.008 15.43 0.143 I LCO 57262.251 15.34 0.193 I LCO 57264.083 15.83 0.150 I LCO 57268.052 15.72 0.129 I LCO 57269.908 15.71 0.133 I LCO 57271.882 15.70 0.143 I LCO 57273.924 15.80 0.148 I LCO 57276.874 15.80 0.142 I LCO 57279.895 15.81 0.140 I LCO 57282.476 15.85 0.133 I LCO 57291.383 15.98 0.154 I LCO 57294.522 15.99 0.141 I LCO 57299.495 15.80 0.139 I LCO 57302.481 15.95 0.200 I LCO 57304.921 15.86 0.262 I LCO 57307.106 15.93 0.126 I LCO 57307.140 15.96 0.125 I LCO 57316.018 15.98 0.134 I LCO 57319.059 16.05 0.152 I LCO 57321.796 16.04 0.116 I LCO 57327.133 16.03 0.126 I LCO 57329.849 15.88 0.128 I LCO 57332.796 16.02 0.159 I LCO 57335.821 16.06 0.157 I LCO 57346.424 15.83 0.140 I LCO 57349.419 15.87 0.133 I LCO 57355.039 15.92 0.121 I LCO 57249.533 16.01 0.053 V LCO 57252.004 16.06 0.055 V LCO

(cont’d) Table A.5. Follow-up Photometric Measurements of ASASSN-15oi.

202 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57256.429 16.1 0.053 V LCO 57258.009 16.3 0.225 V LCO 57262.252 16.52 0.144 V LCO 57266.065 16.55 0.112 V LCO 57268.053 16.63 0.047 V LCO 57269.909 16.76 0.050 V LCO 57271.883 16.79 0.062 V LCO 57273.926 16.83 0.058 V LCO 57276.875 16.93 0.057 V LCO 57278.172 16.94 0.062 V LCO 57282.475 16.98 0.050 V LCO 57285.424 16.9 0.130 V LCO 57291.382 17.03 0.081 V LCO 57294.521 17.13 0.074 V LCO 57304.919 17.22 0.074 V LCO 57307.104 17.18 0.047 V LCO 57307.138 17.18 0.056 V LCO 57312.170 17.2 0.068 V LCO 57315.877 17.28 0.082 V LCO 57315.894 17.14 0.073 V LCO 57316.016 17.24 0.075 V LCO 57319.057 17.23 0.059 V LCO 57321.794 17.14 0.077 V LCO 57327.129 17.26 0.045 V LCO 57329.844 17.25 0.055 V LCO 57332.791 17.28 0.051 V LCO 57335.816 17.29 0.065 V LCO 57344.415 17.22 0.060 V LCO 57346.420 17.17 0.066 V LCO 57349.423 17.32 0.056 V LCO 57355.035 17.25 0.052 V LCO 57259.334 16.12 0.071 V Swift

(cont’d)

203 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57264.392 16.42 0.071 V Swift 57268.788 16.88 0.1 V Swift 57272.514 16.86 0.091 V Swift 57276.096 16.98 0.1 V Swift 57280.688 17.03 0.091 V Swift 57285.415 16.99 0.17 V Swift 57291.590 17.01 0.13 V Swift 57294.588 17.06 0.13 V Swift 57300.236 17.47 0.16 V Swift 57303.628 17.29 0.15 V Swift 57306.494 17.08 0.12 V Swift 57309.413 17.54 0.22 V Swift 57312.137 17.09 0.13 V Swift 57318.671 17.28 0.25 V Swift 57324.051 17.82 0.36 V Swift 57327.961 17.16 0.15 V Swift 57330.420 17.26 0.25 V Swift 57333.480 17.30 0.2 V Swift 57336.141 17.61 0.22 V Swift 57339.334 16.84 0.21 V Swift 57342.069 16.88 0.18 V Swift 57345.793 17.12 0.16 V Swift 57348.384 17.23 0.16 V Swift 57249.531 16.09 0.055 B LCO 57252.001 16.2 0.075 B LCO 57256.426 16.34 0.042 B LCO 57258.007 16.44 0.065 B LCO 57262.250 16.52 0.082 B LCO 57264.081 16.58 0.062 B LCO 57266.062 16.87 0.063 B LCO 57268.051 16.99 0.068 B LCO 57269.906 17.05 0.060 B LCO

(cont’d)

204 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57271.880 17.19 0.074 B LCO 57273.923 17.38 0.067 B LCO 57276.872 17.26 0.069 B LCO 57278.169 17.42 0.076 B LCO 57282.474 17.47 0.065 B LCO 57285.423 17.62 0.083 B LCO 57291.381 17.7 0.115 B LCO 57294.520 17.58 0.242 B LCO 57295.130 17.67 0.085 B LCO 57299.493 17.95 0.098 B LCO 57304.918 17.59 0.086 B LCO 57307.137 17.93 0.067 B LCO 57312.168 17.98 0.077 B LCO 57315.873 18.27 0.130 B LCO 57316.015 18.09 0.086 B LCO 57319.055 18.21 0.108 B LCO 57321.792 18.45 0.179 B LCO 57327.124 18.12 0.058 B LCO 57329.839 18.14 0.049 B LCO 57332.786 18.31 0.099 B LCO 57335.812 18.18 0.102 B LCO 57344.411 18.25 0.077 B LCO 57346.415 18.11 0.131 B LCO 57349.414 18.27 0.076 B LCO 57353.420 18.24 0.064 B LCO 57355.030 18.35 0.071 B LCO 57259.329 16.33 0.045 B Swift 57264.384 16.73 0.045 B Swift 57268.724 16.95 0.054 B Swift 57272.507 17.26 0.063 B Swift 57276.089 17.33 0.063 B Swift 57280.679 17.46 0.063 B Swift

(cont’d)

205 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57285.413 17.57 0.122 B Swift 57291.388 17.67 0.092 B Swift 57294.582 17.68 0.092 B Swift 57300.230 17.71 0.092 B Swift 57303.622 17.80 0.102 B Swift 57306.487 17.68 0.092 B Swift 57309.409 17.80 0.132 B Swift 57312.132 17.68 0.102 B Swift 57318.668 18.13 0.181 B Swift 57324.049 17.91 0.171 B Swift 57327.911 18.44 0.231 B Swift 57330.418 18.08 0.221 B Swift 57333.414 18.08 0.211 B Swift 57336.135 18.17 0.181 B Swift 57342.066 17.84 0.171 B Swift 57345.788 18.26 0.171 B Swift 57348.378 18.27 0.171 B Swift 57259.328 15.25 0.045 U Swift 57264.383 15.69 0.045 U Swift 57268.722 15.99 0.045 U Swift 57272.506 16.19 0.054 U Swift 57276.088 16.51 0.063 U Swift 57280.678 16.73 0.063 U Swift 57285.412 16.99 0.122 U Swift 57291.387 16.96 0.082 U Swift 57294.581 17.26 0.102 U Swift 57300.229 17.36 0.102 U Swift 57303.621 17.47 0.112 U Swift 57306.486 17.45 0.102 U Swift 57309.408 17.67 0.161 U Swift 57312.131 17.79 0.151 U Swift 57318.667 17.46 0.161 U Swift

(cont’d)

206 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57324.048 17.75 0.211 U Swift 57327.910 17.68 0.181 U Swift 57330.418 17.68 0.231 U Swift 57333.413 17.79 0.221 U Swift 57336.134 17.47 0.141 U Swift 57339.331 17.81 0.271 U Swift 57342.066 17.84 0.231 U Swift 57345.787 18.08 0.201 U Swift 57348.377 17.82 0.161 U Swift 57259.325 14.60 0.042 UV W 1 Swift 57264.379 15.02 0.042 UV W 1 Swift 57268.719 15.30 0.05 UV W 1 Swift 57272.503 15.55 0.05 UV W 1 Swift 57276.085 15.89 0.058 UV W 1 Swift 57280.675 16.07 0.058 UV W 1 Swift 57285.411 16.24 0.076 UV W 1 Swift 57291.384 16.52 0.067 UV W 1 Swift 57294.579 16.66 0.067 UV W 1 Swift 57297.048 16.91 0.076 UV W 1 Swift 57300.226 16.91 0.076 UV W 1 Swift 57303.618 17.07 0.076 UV W 1 Swift 57306.483 17.14 0.076 UV W 1 Swift 57309.407 17.16 0.104 UV W 1 Swift 57312.128 17.41 0.095 UV W 1 Swift 57318.666 17.55 0.133 UV W 1 Swift 57324.047 17.98 0.192 UV W 1 Swift 57327.909 17.68 0.143 UV W 1 Swift 57330.416 17.63 0.163 UV W 1 Swift 57333.412 17.60 0.153 UV W 1 Swift 57336.132 17.89 0.133 UV W 1 Swift 57339.330 17.72 0.192 UV W 1 Swift 57342.064 18.06 0.202 UV W 1 Swift

(cont’d)

207 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57345.784 18.26 0.173 UV W 1 Swift 57348.374 18.11 0.143 UV W 1 Swift 57259.335 14.44 0.042 UV M2 Swift 57264.394 14.87 0.042 UV M2 Swift 57268.789 15.21 0.05 UV M2 Swift 57272.515 15.36 0.05 UV M2 Swift 57276.098 15.63 0.05 UV M2 Swift 57280.689 15.79 0.05 UV M2 Swift 57285.416 15.97 0.067 UV M2 Swift 57291.591 16.18 0.05 UV M2 Swift 57294.589 16.38 0.058 UV M2 Swift 57300.237 16.67 0.058 UV M2 Swift 57303.629 16.81 0.058 UV M2 Swift 57306.495 16.93 0.058 UV M2 Swift 57309.413 17.00 0.076 UV M2 Swift 57312.138 17.07 0.067 UV M2 Swift 57318.869 17.40 0.182 UV M2 Swift 57324.052 17.59 0.114 UV M2 Swift 57327.962 17.56 0.076 UV M2 Swift 57330.421 17.41 0.114 UV M2 Swift 57333.481 17.86 0.104 UV M2 Swift 57336.142 17.82 0.085 UV M2 Swift 57339.334 17.84 0.143 UV M2 Swift 57342.069 17.84 0.124 UV M2 Swift 57345.794 17.94 0.242 UV M2 Swift 57348.385 17.92 0.095 UV M2 Swift 57257.794 14.77 0.042 UV W 2 Swift 57259.330 14.83 0.042 UV W 2 Swift 57268.725 15.21 0.076 UV W 2 Swift 57272.508 15.48 0.042 UV W 2 Swift 57276.091 15.68 0.042 UV W 2 Swift 57280.681 15.91 0.042 UV W 2 Swift

(cont’d)

208 Table A.5—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

57285.413 16.01 0.058 UV W 2 Swift 57291.389 16.12 0.05 UV W 2 Swift 57294.583 16.17 0.05 UV W 2 Swift 57300.231 16.35 0.05 UV W 2 Swift 57303.623 16.42 0.058 UV W 2 Swift 57306.488 16.57 0.058 UV W 2 Swift 57309.410 16.56 0.067 UV W 2 Swift 57312.133 16.79 0.058 UV W 2 Swift 57318.668 16.94 0.085 UV W 2 Swift 57324.049 17.16 0.095 UV W 2 Swift 57327.911 16.94 0.143 UV W 2 Swift 57330.418 17.42 0.114 UV W 2 Swift 57336.136 17.29 0.076 UV W 2 Swift 57339.332 17.47 0.124 UV W 2 Swift 57342.067 17.39 0.104 UV W 2 Swift 57345.789 17.51 0.085 UV W 2 Swift 57348.379 17.60 0.085 UV W 2 Swift

aMagnitudes are presented in the natural system for each ﬁlter: ugriz magnitudes are in the AB system, while Swift ﬁlter magnitudes are in the Vega system.

209 Magnitude MJD Magnitudea Uncertainty Filter Telescope

56569.738 17.212 0.045 i LCO 56582.811 17.246 0.040 i LCO 56582.821 17.245 0.043 i LCO 56600.058 17.751 0.058 i LCO 56617.036 17.936 0.069 i LCO 56624.048 18.231 0.097 i LCO 56640.093 19.076 0.232 i LCO 56641.115 18.832 0.179 i LCO 56569.735 17.135 0.023 r LCO 56582.808 17.451 0.026 r LCO 56582.818 17.429 0.025 r LCO 56600.055 17.762 0.027 r LCO 56617.033 17.899 0.031 r LCO 56624.045 18.091 0.038 r LCO 56640.090 18.871 0.102 r LCO 56641.112 19.137 0.124 r LCO 56538.950 16.916 0.220 V Swift 56541.680 16.660 0.174 V Swift 56543.414 16.834 0.194 V Swift 56550.293 17.372 0.617 V Swift 56555.621 16.916 0.209 V Swift 56559.501 17.004 0.239 V Swift 56563.831 17.227 0.278 V Swift 56568.102 16.818 0.212 V Swift 56572.236 17.060 0.265 V Swift 56576.440 17.509 0.381 V Swift 56581.723 17.183 0.297 V Swift 56584.448 17.480 0.371 V Swift 56569.733 18.037 0.065 g LCO 56582.816 18.362 0.051 g LCO 56600.053 18.991 0.070 g LCO 56624.042 19.961 0.182 g LCO

(cont’d) Table A.6. Follow-up Photometric Measurements of ASASSN-13co.

210 Table A.6—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56641.110 20.546 0.388 g LCO 56538.945 16.944 0.104 B Swift 56541.674 17.129 0.114 B Swift 56543.408 17.129 0.114 B Swift 56547.479 17.141 0.134 B Swift 56550.289 17.338 0.149 B Swift 56555.615 17.444 0.142 B Swift 56559.496 17.755 0.187 B Swift 56563.825 18.017 0.238 B Swift 56568.099 18.405 0.360 B Swift 56572.232 18.607 0.424 B Swift 56576.437 18.472 0.389 B Swift 56581.719 >18.168 — B Swift 56584.444 >17.839 — B Swift 56538.944 16.067 0.083 U Swift 56541.673 16.325 0.097 U Swift 56543.407 16.434 0.104 U Swift 56547.478 16.972 0.156 U Swift 56550.289 17.389 0.243 U Swift 56555.614 >17.478 — U Swift 56559.495 >17.596 — U Swift 56563.824 >18.504 — U Swift 56568.098 >17.844 — U Swift 56572.231 >18.270 — U Swift 56576.436 >19.035 — U Swift 56581.718 >18.552 — U Swift 56584.444 >19.605 — U Swift 56538.942 16.513 0.101 UV W 1 Swift 56541.669 16.935 0.119 UV W 1 Swift 56543.404 17.362 0.159 UV W 1 Swift 56547.476 18.133 0.295 UV W 1 Swift 56550.287 >17.868 — UV W 1 Swift

(cont’d)

211 Table A.6—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56555.611 >18.681 — UV W 1 Swift 56559.492 >18.652 — UV W 1 Swift 56563.821 >18.393 — UV W 1 Swift 56568.096 >18.488 — UV W 1 Swift 56572.229 >18.388 — UV W 1 Swift 56576.434 >18.900 — UV W 1 Swift 56581.716 >18.501 — UV W 1 Swift 56584.442 >18.133 — UV W 1 Swift 56538.951 16.915 0.157 UV M2 Swift 56541.681 17.604 0.129 UV M2 Swift 56543.415 17.902 0.165 UV M2 Swift 56550.756 17.961 0.563 UV M2 Swift 56555.622 >18.474 — UV M2 Swift 56559.502 >18.762 — UV M2 Swift 56563.832 >18.434 — UV M2 Swift 56568.103 >18.636 — UV M2 Swift 56572.237 >18.430 — UV M2 Swift 56576.441 >18.400 — UV M2 Swift 56581.723 >18.773 — UV M2 Swift 56584.449 >18.676 — UV M2 Swift 56538.946 17.407 0.122 UV W 2 Swift 56541.675 17.719 0.151 UV W 2 Swift 56543.409 18.159 0.212 UV W 2 Swift 56550.290 >18.472 — UV W 2 Swift 56555.616 >19.275 — UV W 2 Swift 56559.497 >19.188 — UV W 2 Swift 56563.826 >18.675 — UV W 2 Swift 56568.099 >18.709 — UV W 2 Swift 56572.233 >18.735 — UV W 2 Swift 56576.437 >18.833 — UV W 2 Swift

(cont’d)

212 Table A.6—Continued

Magnitude MJD Magnitudea Uncertainty Filter Telescope

56581.720 >18.787 — UV W 2 Swift 56584.445 >18.534 — UV W 2 Swift

aMagnitudes are presented in the natural system for each ﬁlter: ugriz magnitudes are in the AB system, while Swift ﬁlter magnitudes are in the Vega system. 3-sigma upper limits are given for epochs with no detection.

213