Title Study of Dust-Torus Properties Around Supermassive Black Holes

Study of Dust-Torus Properties around Supermassive Black Title Holes( Dissertation_全文 )

Author(s) Ichikawa, Kohei

Citation 京都大学

Issue Date 2015-03-23

URL https://doi.org/10.14989/doctor.k18796

Right 許諾条件により本文は2015/10/01に公開

Type Thesis or Dissertation

Textversion ETD

Kyoto University Study of Dust-Torus Properties around Supermassive Black Holes

Kohei Ichikawa Department of Astronomy Kyoto University

A thesis submitted for the degree of Doctor of Philosophy

January, 2015 Abstract

The apparent ubiquity of supermassive black holes (SMBHs) in the spheroidal components of present-day galaxies, and the correlation between the masses of SMBHs and spheroidal components indicates that active galactic nuclei (AGN; SMBH-driven activity) and star-formation (SF; plausibly the progenitors of spheroids) are physically connected. While AGN present a variety of observational characteristics, the uniﬁed model of AGN proposes the ubiquitous presence of a dusty, obscuring, and geometrically thick “torus” around the central engine, and that all AGN are fundamentally same (Antonucci, 1993). Considering the tori could be the media providing the mass to the central engine, knowing the torus structure such as the dust mass and the size is essential to investigate the physical link between SMBHs and the host galaxies. However, the dust structures of the tori are not well understood due to their small scale (r < 10 pc). Understanding torus models enable us to quantitatively derive the precise torus infrared emission and well decompose the infrared spectra into SF and dust-torus emission. One of the crucial way to proceed torus studies are mid-infrared (3–20 µm) observations because they trace well the dust-torus emission.

We first investigate the mid-infrared (MIR) properties of a nearly complete sample of local AGN detected in the Swift/Burst Alert Telescope (BAT) all-sky hard X- ray (14-195 keV) survey, based on the cross correlation with the AKARI, IRAS, and WISE infrared survey catalogs. Out of 135 non-blazar AGN in the Swift/BAT nine-month catalog, we obtain the MIR photometric data for 128 sources. We find good correlation between their hard X-ray and MIR luminosities over three orders of magnitude (42 < log λLλ(9, 18µm) < 45). In addition, both X-ray unabsorbed and absorbed AGN follow the same correlation, implying isotropic infrared emission, as expected in clumpy dust tori rather than homogeneous ones. We find excess signals around 9 µm in the averaged infrared spectral energy distribution from heavily obscured “new type” AGN with small scattering fractions in the X-ray spectra. This could be attributed to the polycyclic aromatic hydrocarbon emission feature, suggesting that their host galaxies have strong starburst activities. Next, we apply the clumpy torus model to the observed AGN dust-torus spectra. Against the fact that the observations generally support the unified model, there is the question why some, but not all, type-2 AGN do not show any observational signs of polarized broad line (PBL) regions. To understand the role of the obscuration by the torus in type-2 AGN and the detectability of PBLs, knowing the torus geometry/morphology and properties is essential. These sources are observed with high spatial resolution (∼ 0.3–0.7 arcsec) mid-infrared N band spectroscopy, Q band imaging, and nuclear near-infrared photometries obtained from the literature. Combining these high spatial resolution infrared observations and clumpy torus models enables us to put tight constraints on the torus properties and geometry. We divide the sample into three types according to the broad line region (BLR) properties; type-1s, type-2s with scattered or hidden broad line region (HBLR) previously observed, and type-2s without any published HBLR signature (NHBLR). Comparing the torus parameters gives us the first quantitative torus geometrical view for each subgroup. We find that NHBLR AGN have smaller torus opening angle, larger number of torus clumps, and larger covering factor than those of HBLR AGN. This suggests that the chance to observe scattered (polarized) flux from the BLR in NHBLR could be reduced by the dual effects of (a) less scattering medium due to the reduced scattering volume with small torus opening angle and (b) increased torus obscuration between the observer and the scattering region. These effects give a reasonable explanation for the lack of observed HBLR in some type-2 AGN.

Finally, we search heavily obscured (=buried) AGN using AKARI /IRC 2.5–5.0 µm spectra. Many studies imply that the cosmic evolution of obscured AGN are somehow coupled to the star formation activity, which peaks at z ∼ 2. Therefore, complete AGN surveys covering heavily obscured (=buried) AGN including Compton- thick populations are crucial to understand the cosmic history hidden by dust/gas obscuration. Candidates of galaxies that host such buried AGN are infrared galax- 10 ies, defined as those having infrared luminosity (LIR) with LIR > 10 L . We conduct AKARI 2.5–5.0 µm spectroscopy of such infrared galaxies without any signs of AGN in the optical. Applying our AGN diagnostics to the AKARI spectra by decomposing the spectra into stellar and AGN dust-torus components, we find that both the fraction and the energy contribution of buried AGN increase with 10 13 infrared luminosity from 10 L to 10 L . However, the energy contribution from AGN in the total infrared luminosity is only ∼7% in LIRGs and ∼20% in ULIRGs, suggesting that the majority of the infrared luminosity originates from starburst activity, not AGN. This simple method will have a great advantage in the era of JWST, because we can apply our method to high-z galaxies by observing the rest 2.5–5.0 µm band (5.0–10.0 µm at z = 1 and 7.5–15.0 µm at z = 2) with JWST /MIRI (5–28 µm). Other methods based on the Spitzer bandpass (5–35 µm) would have difficulty studying high-z objects; for instance, the observed spectral range corresponds to 15–105 µm at z = 2, which only the SPICA mission can cover. Thus, we have essential method for studying infrared galaxies in the distant universe even before the launch of SPICA (from 2028–). Acknowledgements

During my PhD course, my study on astronomy and daily life have been supported by many people. Even after three years of PhD course, I still do not have enough skills and abilities to express my feeling of acknowledgements into words to the everyone I would like to thank.

First of all, I sincerely would like to thank my supervisor, Dr. Yoshihiro Ueda, for all of your efforts and advices for the past five years. You encouraged me and gave me very accurate scientific comments everytime. All these things allowed me to have a wide range of astronomical viewpoints and now I can stand on the starting point as a professional astronomer. Thank you for your help and patience. I am very grateful to my collaboratos, for giving me many productive suggestions through the collaborations. I would like to thank Dr. Masatoshi Imanishi for giving me the great opportunity to enter the field of infrared astronomy. Most of my knowledge and observational experiences on infrared astronomy first comes from you. Thanks to your generous support on my research, now I can stand on the start point as a professional infrared astronomer. My hearfelt thanks to Dr. Christpher Packham. You allowed me to expand my study on AGN with infrared observations and introduced me brightest AGN people worldwide. Through the collaboration with you, I learned how to collaborate with people and how to improve the study through collaborations with bright people all over the world. During the stay in Texas, you always gave me exciting opportunities on astronomy and gave me many tips on how to live in the U.S. I am grateful to Dr. Claudio Ricci through the collaboration during your stay in Kyoto. The discussion on AGN with you was one of my precious time to deepen my understanding on AGN. Your sincere attitude on the study of astronomy was very impressive to me and promoted my motivation on astronomy. I also would like to acknowledge Almudena Alonso-Herrero, Cristina Ramos Almeida, AndrésAsensio Ramos, Omaira González-Mart´ın,Enrique Lopez- Rodriguez for your continuous assistance and advices on my research and infrared astronomical tools including BayesClumpy and Redcan. I would like to express my thanks to Tanio D´ıaz-Santos, Moshe Elitzur, Sebastian F. Hönig,Nancy A. Levenson, Rachel E. Mason, Eric S. Perlman, Crystal D. Alsip, Takao Nakagawa, Mai Shirahata, Hidehiro Kaneda, Shinki Oyabu, Yuichi Terashima, Poshak Gandhi, and Keiko Matsuta for your irreplaceable advices and fruitful suggestions on my researches.

I would like to also express my sincere appreciations to all members of our scientiﬁc group. Dr. Masaaki Hayashida, Dr. Marko Stalevski, Dr. Murray Brightman, Dr. Kenta Matsuoka, Dr. Satoshi Eguchi, Dr. Kazuo Hiroi, Dr. Fumie Tazaki, Ms. Megumi Shidatsu, Mr. Ryosuke Sato, Mr. Taiki Kawamuro, Mr. Takafumi Hori, and Mr. Shota Mizuno. I was very happy and felt proud that I have spent very important days with you during my PhD course.

My heartfelt thanks also to all the people in the department of Astronomy, Kyoto University. My experiences during my PhD course has been wonderful time thanks to all of you. My skill of Mahjong has been improved signiﬁcantly thanks to over 100 matches with you.

I thank my family for their long time support over 26 years. My hard-working parents have sacriﬁced their lives for me and my brothers with full of love and care. I love you so much, and I would not have made it this far without you. My brothers have been my friends all of my life since I was a baby. Finally, I woud like to thank Atsuko. You are my dearest partner. You have been supportive and caring even when I was in a hard time. I would never been able to be here today without you. I look forward to sharing the future with you. Contents

List of Figures vii

List of Tables ix

1 General Introduction on Active Galactic Nuclei 1 1.1 Brief history of Active Galactic Nuclei ...... 1 1.2 Components in AGN ...... 3 1.2.1 Supermassive black hole ...... 3 1.2.2 Accretion Disk ...... 3 1.2.3 Broad Line Region; BLR ...... 4 1.2.4 Narrow Line Region ...... 4 1.3 Unified Model of AGN ...... 5 1.4 Torus Models on the Market ...... 6 1.5 Application of Clumpy Torus Models to the Observed Data ...... 10 1.5.1 The difference between type-1 and type-2 AGN ...... 10 1.5.2 The difference between Hidden BLR AGN and Non-Hidden BLR AGN . . 11 1.5.3 Luminosity dependence of AGN torus ...... 12 1.6 Co-Evolution of SMBHs and Their Host galaxies ...... 13 1.7 Obscured AGN Survey ...... 14 1.8 Motivation and Strategy of This Study ...... 16

2 Mid- and Far-infrared Properties of a Complete Sample of Local Active Galactic Nuclei 19 2.1 Introduction ...... 19 2.2 Sample ...... 21 2.2.1 Swift/BAT Hard X-ray Catalog ...... 21 2.2.2 Infrared Catalogs ...... 21 2.2.2.1 AKARI Point Source Catalogs ...... 21 2.2.2.2 IRAS Catalogs ...... 21

iii CONTENTS

2.2.2.3 WISE All-Sky Catalog ...... 22 2.2.3 Cross Correlation of Swift/BAT AGN with the IR Catalogs ...... 22 2.2.4 AGN Type ...... 24 2.2.5 Luminosity Correlation between AKARI, IRAS, and WISE Data . . . . . 24 2.2.6 Basic Properties of the Sample ...... 26 2.3 Results and Discussion ...... 32 2.3.1 Correlation between the Infrared and Hard X-ray Luminosities ...... 32 2.3.2 Averaged IR Spectral Energy Distribution ...... 37 2.4 Summary and Conclusions ...... 39

3 Clumpy torus modeling of polarized and non-polarized BLR AGN 41 3.1 Introduction ...... 41 3.2 Observations ...... 43 3.2.1 The Sample ...... 43 3.2.2 New Observations ...... 44 3.2.3 Published Data from the Literature ...... 44 3.2.4 Subsample ...... 47 3.3 Application of Torus Model ...... 47 3.3.1 Clumpy Torus Model ...... 47 3.3.2 Other Important Torus Parameters ...... 49 3.3.3 BayesClumpy and Modeling Details ...... 49 3.4 Results and Discussions ...... 50 3.4.1 Infrared SEDs with BayesClumpy Fitting ...... 50 3.4.2 General Torus Properties of Total Sample ...... 54 3.4.3 Distribution of Torus Parameters ...... 55 3.4.4 Distribution of Covering Factor ...... 57 3.4.5 Torus Morphological Diﬀerences between HBLR and NHBLR ...... 58 3.4.6 Inclination Angle Eﬀect on Detectability of HBLR in Type-2 AGN . . . . 59 3.5 Conclusions ...... 60

4 AKARI IRC 2.5–5 µm Spectroscopy of Infrared Galaxies over a Wide Lumi- nosity Range 63 4.1 Introduction ...... 63 4.2 Targets ...... 65 4.3 Observations and Data Reduction ...... 66 4.4 AGN/SB Spectral Decomposition and Buried AGN Diagnostics ...... 67

iv CONTENTS

4.4.1 AGN-heated dust component ...... 68 4.4.2 Stellar and Starburst Component ...... 69 4.4.3 Emission and Absorption lines ...... 69 4.4.4 Total model spectra ...... 70 4.5 Results and Discussions ...... 73 4.5.1 AKARI spectra and AGN diagnostics ...... 73 4.5.2 Buried AGN contribution as a function of infrared luminosity ...... 76 4.5.3 Luminosity contribution of buried AGN to the total infrared luminosity . 77 4.5.4 Evolutionary track of IR galaxies ...... 80 4.6 Conclusions ...... 81

5 Conclusion and Outlook 83 5.1 Conclusion ...... 83 5.1.1 Constraints of the torus models on the market ...... 83 5.1.2 Clumpy torus model application to the observed data ...... 83 5.1.3 Obscured AGN search obtained from infrared observations ...... 84 5.2 Future Prospects for AGN Studies through Dusty Tori ...... 84 5.2.1 Unveiling the torus properties of Low-Luminosity AGN (LLAGN) . . . . 85 5.2.2 Origin of dust-free AGN (DFAGN) in the local universe ...... 86 5.2.3 ALMA observations of those AGN populations ...... 87

References 89

v CONTENTS

vi List of Figures

1.1 High spatial image of ionization cone ...... 5 1.2 Schematic image of AGN uniﬁed model ...... 6 1.3 Half-light radii of AGN tori as a function of UV luminosity ...... 7 1.4 Smooth torus modeled and Observed spectra of AGN ...... 7 1.5 Flux distribution of the physical model with MIR interferometric observations . . 8 1.6 Schematic image of clumpy torus model ...... 9 1.7 Inclination angle dependence of the SEDs of the clumpy torus ...... 9 1.8 Distribution of the covering factor and escape probability of type-1 and type-2 AGN...... 10

1.9 MBH–σ relation for AGN (blue circles) and quiescent galaxies (red stars). . . . . 13 1.10 Cosmic X-ray Background spectrum calculated by Ueda et al. (2014)...... 15

2.1 Positional diﬀerence between Swift/BAT and AKARI sources ...... 23 2.2 Infrared Correlation plots ...... 25 2.3 Redshift Distribution ...... 26 2.4 Redshift Distribution ...... 28 2.5 Mid-Infrared and Hard X-ray luminosity Correlations ...... 32 2.6 Histgram of mid-infrared to hard X-ray luminosity ratio ...... 34

2.7 MIR to hard X-ray luminosity ratio as a function of NH ...... 35 2.8 Averaged IR SED ...... 38

3.1 Model fits for type-1 AGN...... 51 3.2 Model fits for HBLR AGN...... 52 3.3 Model fits for NHBLR AGN...... 52 3.4 Posterior Distribution of each parameter...... 55 3.5 Posterior Distribution of derived parameter...... 56 3.6 Schematic illustration of the torus geometry ...... 58

vii LIST OF FIGURES

3.7 Color-color plot obtained from IRAS and BayesClumpy ...... 59

4.1 Redshift distribution of our AKARI sample ...... 65 4.2 AKARI 2.5–5.0 µm spectra ...... 72 4.3 Equivalent width of 3.3 µm PAH vs. dust temperature ...... 75 4.4 Buried AGN fraction as a function of infrared luminosity...... 76

4.5 Ratio of L3.3PAH/LIR as a function of infrared luminosity...... 77 4.6 Energy contribution of buried AGN to the total infrared luminosity as a function of infrared luminosity...... 78 4.7 Comoving infrared luminosity density as a function of redshift...... 79 4.8 Evolutional Track of AGN...... 81

5.1 Obscured AGN fraction (=covering factor) as a function of AGN luminosity. . . 85 5.2 Luminosity correlation between 9 µm and hard X-ray band...... 87

viii List of Tables

2.1 Infrared and X-ray Properties of the AGN in the Swift/BAT 9-month AGN catalog 29 2.1 Infrared and X-ray Properties of the AGN in the Swift/BAT 9-month AGN catalog 30 2.1 Infrared and X-ray Properties of the AGN in the Swift/BAT 9-month AGN catalog 31 2.2 Correlation parameters between MIR and the hard X-ray sample ...... 33

2.3 Average and standard deviation of log(λLλ(9, 18µm)/LHX)...... 34

3.1 Properties of the Sample ...... 43 3.2 List of Photometry ...... 46 3.3 Free Parameters of the BayesClumpy ...... 49 3.4 Fitted torus model parameters from SED + spectroscopy data ...... 53 3.5 Torus model parameters from the global posterior distributions ...... 54 3.6 Results of KLD test for each parameter among each subgroup ...... 55

4.1 Basic Information of Our Sample ...... 68 4.2 Fitting Properties and AGN Signs Obtained from the AKARI 2.5–5.0 µm Spec- troscopy ...... 70 4.3 Observed Properties Obtained from the AKARI 2.5–5.0 µm...... 74

ix LIST OF TABLES

x 1

General Introduction on Active Galactic Nuclei

1.1 Brief history of Active Galactic Nuclei

The history of Active Galactic Nuclei (AGN) started from the beginning of twentieth century. The first optical spectra of the AGN were undertaken by Fath (1909) at Lick Observatory. At the time, a major issue was to know whether spiral galaxies were relatively nearby gaseous objects or very distant star clusters. The author found that most of the sources show continuous spectra with stellar absorption lines, suggesting an unresolved composite of stars. However, in the case of NGC 1068, the author noted the presence of strong emission lines in the spectrum. Later Slipher (1917) confirmed the same properties of NGC 1068 with higher spectral resolution spctrum obtained in 1913 at Lowell Observatory, and showed that these lines have widths of several 100 km/s. Hubble (1926) followed the observations of the same kind of galaxies, and reported that NGC 4051 and NGC 4151 also show the strong emission lines as well as NGC 1068. The first systematic study of galaxies showing strong emission lines was conducted by Seyfert (1943). The author selected a population of galaxies that show strong central surface brightness, and showed that the optical spectra of those galaxies have both broad emission lines (up to ∼ 8500 km/s) and narrow forbidden emission lines. However, until NGC 1068 and NGC 1275 were detected as radio sources, the study of Seyfert (1943) and the topic of those AGN had received few attentions. The strong impetus for the study of AGN starts from the discovery of radio sources in the extrasolar system (Jansky, 1933). After the World War II, several radio engineer groups joined the study of radio astronomy. Utilizing an interferometer, Bolton & Stanley (1948) confirmed one source called Cyg A has a brightness temperature with > 4 × 106 K at 100 MHz and they concluded the source cannot be a thermal origin. The same kind of results have been reported

1 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

by Bolton (1948) for more six sources. Growing number of studies on position determination and optical identification of those radio sources were conducted in 1950s. Smith (1951) reported accurate positions of those radio sources. Those data first enable Baade & Minkowski (1954) to derive detailed optical spectroscopic studies of those radio sources. Using Hα and some other forbidden lines, they first reported that Cyg A is located at a large distance of 31 Mpcs, and the calculated luminosity is ∼ 1043 erg/s in the radio and the optical. Furthermore, using the same method, Minkowski (1960) reported 3C 295 is located at z ∼ 0.46. These distant, stellar like sources came to be known as quasi-stellar objects, or “QSOs”. A detailed study was conducted by Greenstein & Schmidt (1964). They discussed the origin of the redsfhit whether gravitational one or cosmological one. Considering the observed QSOs have clear symmetry line profiles, gravitational redshift model seems unnatural. Futhermore, for an object with the mass of 1M , the observed Hβ flux should have an electron density with ∼ 1019 cm−3, which is well above the critical density of the forbidden lines in the spectra. They concluded that the emission line redhisfts could 9 be originated from cosmological redshifts, and QSO requires a mass of over 10 M and it is located within a Schwarzschild radius of ∼ 10−4 pc, suggesting the existence of supermassive black holes in the center (e.g., Lynden-Bell, 1969). The discovery of a large population of radio-quiet AGN were first reported by Sandage (1965). They found many radio-quiet AGN using UBV color-color diagram based on ultraviolet excess originated from accretion disks of those AGN. Kristian (1973) showed that there are diffuse source surrounding the central compact objects, suggesting the presence of host galaxies. After the improvement of optical spectroscopic observations, emission lines of AGN have been investigated. The basic properties of the region of gas emotion the narrow forbidden emission lines were first established. Using the line ratio of [Sii] and [Oiii], Woltjer (1959) 4 −3 4 calculated the electron density ne = 10 cm and the temperature T ∼ 2 × 10 K. The physical size of those narrow emission region was resolved with an order of 1 kpc (Walker, 5 1968). Oke & Sargent (1968) derived a mass of those narrow emission regions with ∼ 10 M . There was a question that the origins of broad emission lines and narrow emission lines. AGN also show broad lines with v > 3000 Km/s as well as narrow forbidden lines. Woltjer (1959) proposed a separate region of fast-moving, gravitationally bound gas emitting the broad Balmer lines (so called “Broad Line Regions; BLR”), and the region of larger, slower moving, less dense gas emitting the narrow Balmer/forbidden emission lines (“Narrow Line Regions; NLR”). In this stage, many features had been already reported currently recognizable idea on AGN. AGN consists of a supermassive black hole, accretion disk, broad emission line regions, and

2 1.2 Components in AGN

narrow, forbidden emission line regions. In the following, I illustrate the several important components that constitute AGN, then describe the uniﬁed model of AGN, and the co-evolution of the galaxies and the nuclei to connect the motivation of this dissertation.

1.2 Components in AGN

AGN are fundamentally powered by gravitational accretion onto supermassive black holes. Based on the observations of AGN, they are thought to consist of several components: the supermassive black hole, the accretion disk, the broad line region, the narrow line region.

1.2.1 Supermassive black hole

The presence of supermassive black holes (SMBHs) in the center of AGN was ﬁrst theoretically predicted by Lynden-Bell (1969), and following observations conﬁrmed the theory. The upper limit of the size is given from the short time scale variation with less than one hour observed in AGN (suggesting < 0.001 pc).

One of the most common ways to measure the black hole mass (MBH) within AGN is through the observations of broad emission lines. The lines are emitted from the gas rotating around the SMBH and have a velocity broadening of v ∼ 104 km/s. Under the Kepler assumption,

MBH can be derived with rv2 M = (1.1) BH G The last unknown parameter is the distance r, between the place of SMBH and the area of broad emission line regions. The distance r has a clear dependence of central engine luminosity √ L, with r ∝ L, obtained from the optical continuum emission. Combining these relations, the

MBH can be derived with 2 MBH FWHM(Hβ) λLλ(5100) 8 = 8.3 3 44 , (1.2) 10 M 10 km/s 10 erg/s which is estimated by Vestergaard & Peterson (2006).

1.2.2 Accretion Disk

Matter falling onto a SMBH acquires kinetic energy because of its gravitational energy release. Considering the conservation of angular momentum, the matter cannot fall directly onto the SMBH, but forms an accretion disk. The simple case of accretion disks are optically thick, but geometrically thin disks (Shakura & Sunyaev, 1973). In this accretion disk picture, the p matter follows Kepler orbits at any radius with the velocity v ∼ MBH/R, where R denotes

3 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

the distance from SMBH to the matter. Angular momentum is transported outwards through the viscosity. The emission of the accretion disk can be well explained in the form of multi- temperature blackbodies. At the distance R, the disk has a temperature with

T (R) ∝ R−3/4 (1.3)

Considering the temperature has a gradient as a function of R, the emission with maximum tem-

perature Tmax of the accretion disk originates from the region at the distance of Schwarzschild radius, and can be written

1 1 ! 4 1 − ˙ − 4 5 M η 4 M Tmax = 1.4 × 10 7 K. (1.4) 0.1M˙ Edd 0.08 10 M

The typical value of Seyfert AGN has an energy of kTmax = 12 eV, indicating the peak is locate at the UV wavelength. Thus, the disk emission alone cannot account for the observed AGN Xray emission. The most dominant emission in the X-ray is thought to be originated from high temperature electron corona. Zdziarski et al. (1999) mentioned that the X-rays are produced by thermal Comptonization and reprocessed by the material responsible for the seed photons.

1.2.3 Broad Line Region; BLR

The broad emission lines are thought to be originated from the rapidly moving gas in a region close to the SMBH. The typical widths of the lines have an order of v ∼ 103 km/s, with an average of ∼ 5×103 km/s. The broadening is not due to thermal motion. If one assumes the gas 4 temperature T ∼ 10 K, The velocity dispersion of such gas would be v ∼ (kT/mp) ∼ 10 km/s, which is two orders smaller than the typical velocity. Thus, the broadening is due to bulk motions of each gas clouds. The electron density has a lower limit based due to the absence of narrow forbidden lines including [O iii]λ5007 with > 108 cm−3. The strong Lyα and C iv 11 −3 emissions suggest the BLR has ne ∼ 10 cm or even higher. The mass of the BLR MBLR is calculated from C iv luminosity with

MBLR −3 L(C iv) ∼ 10 42 , (1.5) M 10 erg/s

with a mass up to MBLR ∼ 10 M . The physical size RBLR is compact based on the estimation 0.7 from the central AGN luminosity with RBLR ∝ L and the typical size is 0.01 − 0.1 pc for nearby Seyfert AGN.

1.2.4 Narrow Line Region

The narrow emission lines are thought to be originated from the slow moving gas in an extended region. The width of those narrow lines have ∼ 500 km/s. The physical size has a strong relation

4 1.3 Uniﬁed Model of AGN

Figure 1.1: High spatial image of ionization cone - HST image of [Oiii] ionization cone of NGC 1068. The ﬁeld of each panel is 4.0×4.0 arcsec (= 80 × 80 pc in physical scale) (Wilson et al., 2000).

0.5 with the central AGN luminosity with RNLR ∝ L with the typical size of 100 to a few kpc

(see Figure 1.1). The mass of the NLR MNLR can be derived from Hβ luminosity with

MNLR 5 L(Hβ) ∼ 7 × 10 38 . (1.6) M 10 ne 1.3 Uniﬁed Model of AGN

Most of AGN are roughly classified in two types: type-1 and type-2. The difference between two types are based on the observed existence of BLR in the optical spectra. Rowan-Robinson (1977) raised the possibility that the BLR of type-2 AGN was obscured by dust rather than being genuinly absent. The discovery of polarized broad emission lines in the spectra of type-2 AGN (e.g., Antonucci & Miller, 1985; Antonucci, 1993) concreted the idea that AGN are surrounded by a donut-shaped dusty material. They mentioned that electrons in the polar region scatter the emission from the BLR and finally arrives at observers. The electrons in the polar region act as a mirror of BLR. As shown in FIgure 1.2, Krolik & Begelman (1986) called those dusty material “obscuring torus” because a toroidal geometry can explain the angle-dependent, geometrically and optically thick dusty obscuration. The torus dust emits in infrared band due to the heating by the optical-UV photons from the central engine. In the 0.1–100 µm spectra of AGN, one

5 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

can see clear spectral turn-over from the big-blue bump to the infrared bump around 1-3 µm (Edelson & Malkan, 1986). This matches the sublimation temperature of the dust (T ∼ 1500 K) and provides a concrete proof that this feature is originated from the dust.

Figure 1.2: Schematic image of AGN uniﬁed model - The image shows the geometry of the uniﬁed AGN model. The broad line and narrow line regions are shown, as well as the obscuring dusty torus (Urry & Padovani, 1995).

The dust in the torus sublimates when the dust achieves the sublimation temperature heated by the central engine of AGN. The dust sublimate radius can be theoretically written as s r L in = 0.4 AGN . (1.7) pc 1045erg/s

This also matches to the observed dust sublimation radius based on dust reverberation mapping (Suganuma et al., 2006) and NIR interferometry (Kishimoto et al., 2011, ; also see Figure 1.3).

1.4 Torus Models on the Market

In recent years much progress has been made toward understanding the torus structures. Pio- neering works in modeling the torus (Efstathiou & Rowan-Robinson, 1995; Pier & Krolik, 1992, 1993) assume that the torus has uniform and smooth dust density distribution for simplicity, mainly due to the computing power at that time, whereas Krolik & Begelman (1988) already suggested that smooth dust distribution cannot survive within the AGN vicinity. This is because if the velocity dispersion in the center of galaxies due to the thermal motion of smooth

6 1.4 Torus Models on the Market

Figure 1.3: Half-light radii of AGN tori as a function of UV luminosity - The red and green dotted lines are power-law ﬁt to half light radius at 13 µm and 8.5 µm, respectively (Kishimoto et al., 2011). torus dust, then the temperature would achieve over 2000 K, which is too high for the dust to survive. They also proposed that the torus has clumpy structure in order not to be destroyed by the hot gas. Because this temperature issue can e easily solved if the velocity dispersion reﬂects the random1992ApJ...401...99P motion of clouds. In spite of their simple approximation of the “smooth” torus model, encouraging results were reported. One main result which is well explained is the 10 µm silicate feature which depends on the inclination angle and it also matches to the observations at the zeroth order (Hao et al., 2005, 2007; Pier & Krolik, 1993, and Figure 1.4).

Figure 1.4: Smooth torus modeled and Observed spectra of AGN - (Left) A representative AGN torus spectra based on the smooth torus model (Pier & Krolik, 1992). (Right) The observed average spectra of QSOs (red), Seyfert 1s (green), Seyfert 2s (blue), and ULIRGs (black) (Hao et al., 2007).

After the models were proposed, however, many discrepancies between observations and

7 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

the smooth torus models were reported by several authors. First issue is the torus size. As shown in Figures 1.3 and 1.5, recent mid-infrared (MIR) direct imaging observations have well constrained the torus size up to several pc (i.e., Packham et al., 2005; Radomski et al., 2008) and MIR interferometric observations have revealed resolved components whose size has a few pc (Burtscher et al., 2009; Jaﬀe et al., 2004; Raban et al., 2009; Tristram et al., 2007, 2009; Tristram & Schartmann, 2011), while smooth torus models need ∼ 100 pc scale tori for reproducing the infrared (IR) spectral energy distribution (SED).

Figure 1.5: Flux distribution of the physical model with MIR interferometric observations - (Left) Flux distribution of the physical model for the extended component at 11 µm. . (Right) Sketch of the physical model (Tristram et al., 2007).

The second issue is the MIR and X-ray luminosity relations. Considering the geometrical structure of smooth dust models, MIR should emit anisotropically. Their theoretical studies as shown in Figure 1.4 suggest that the MIR luminosity of type-2 AGN drops over one order of magnitude compared to that of type-1 AGN with the same AGN power because the torus itself blocks the inner MIR re-emission from the edge-on view (Levenson et al., 2009). Despite of this theoretical assumption, earlier works suggested that significant differences (only up to a factor of 3) between type-1 and type-2 AGN in the MIR, but with biased, neither complete nor statistically significant number of sample. Therefore, the discussion with complete sample of AGN has been urged. This part will be discussed extensively in Chapter 2. The final issue is the observational events that cannot be explained under the situation of smooth dust models. Smooth torus models assume that the AGN type only depends on the viewing angle, whether or not the dust wall of the torus blocks the broad line emission from the observers’ line of sight. However, some authors reported the transition toward a type-1 AGN before experienced as the type-2 AGN. This event was clearly detected in NGC 7582 (Aretxaga et al., 1999) and possibly

8 1.4 Torus Models on the Market

detected in other AGN (Khachikian & Weedman, 1971; Storchi-Bergmann et al., 1993). The transition from the type-2 to type-1 AGN was also reported in some AGN (Antonucci & Cohen, 1983).

Figure 1.6: Schematic image of clumpy torus model - (Nenkova et al., 2008b).

Figure 1.7: Inclination angle dependence of the SEDs of the clumpy torus - Each line shows the modeled clumpy torus SED at each inclination angle.

In order to solve these issues, and thanks to the improvement of computing power, more physically realistic “clumpy” torus models have been coded by several authors (Hönig& Kishi- moto, 2010; Hönig et al., 2006; Nenkova et al., 2002, 2008a,b; Schartmann et al., 2008, ; see Figures1.6 and 1.7). The first effort of coding the clumpy model was undertaken by Nenkova et al. (2002) and Nenkova et al. (2008a,b). They used 1D radiative transfer to compute the

9 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

torus SED, and suggested that clumpy media can reproduce well the observational characteristics including a silicate feature and the small MIR emission differences between type-1 and type-2 AGN (Levenson et al., 2009). Hönig et al. (2006) and their up-grade studies of Hönig& Kishimoto (2010) developed the clumpy model of Nenkova et al. (2002) by using a 2D radiative transfer code and also took into account various illumination patterns of clumps. They reported silicate features are strongly affected by the original distribution and the density of clumps in the inner region. Schartmann et al. (2008) presented the 3D radiative transfer version of the studies of Hönig et al. (2006) and also demonstrated the results in agreement with Höniget al. (2006). A clumpy torus model by considering the torus inner structure and their inclination angle effect was also reported by Kawaguchi & Mori (2010, 2011), suggesting the strength of near-IR (NIR) emission could be affected by the torus inner structure. Another interesting approach of torus model is the combination of smooth and clumpy torus model called two-phase model in their study (Stalevski et al., 2012). They suggested that their two-phase model can naturally reproduce the observed NIR excess of the torus. In general, these models effectively reproduce NIR to MIR SED with small torus size with up to 10 pc and better explain the actual observations discussed above. This torus model issue is extensively discussed in Chapter 2.

1.5 Application of Clumpy Torus Models to the Observed Data

1.5.1 The diﬀerence between type-1 and type-2 AGN

Figure 1.8: Distribution of the covering factor and escape probability of type-1 and type-2 AGN - The parameter distributions of the torus covering factor (left panels) and escape probability (right panels) for type-1 (top) and type-22 galaxies (bottom) (Ramos Almeida et al., 2011a).

Thanks to the improvement of those clumpy torus models, those models have been applied to

10 1.5 Application of Clumpy Torus Models to the Observed Data

the observed spectra and the SED. The clumpy torus models of Nenkova et al. (2008a,b) with a Bayesian approach has fitted the data to derive the torus physical parameters (BayesClumpy torus models; Asensio Ramos & Ramos Almeida, 2009). In the series of the papers (Ramos Almeida et al., 2009, 2011a), they fitted the NIR to MIR photometric SED of a sample of local type-1 and -2 AGN. They constrained the physical parameters of the sample and found the tentative result that type-2 AGN have a larger torus scale height and more clumps than those of type-1 AGN. This result suggests some properties or evolutional histories of the tori are intrinsically different between type-1 and type-2, which means AGN do not always obey the simple unified model predictions (see Figure 1.8). Alonso-Herrero et al. (2011) expand those previous works by combining the high angular resolution (∼ 000.3–000.4) MIR spectroscopy observations and strongly constrain the torus parameters. They found both of torus geometrical covering factor and scale height depend on the AGN bolometric luminosities. This result may support the receding torus scenario (Lawrence, 1991).

1.5.2 The diﬀerence between Hidden BLR AGN and Non-Hidden BLR AGN

Those studies above cast an idea that AGN unified model is not so simple, but more complicated. The most compelling evidence of the unified model was the detection of polarized broad emission lines (PBLs) in type-2 AGN (e.g., Antonucci & Miller, 1985). Further evidence supporting the unified model comes from infrared (IR) observations in several type-2 AGN which showed the existence of obscured/hidden broad line regions (HBLRs) detectable only with dust penetrating infrared observations (e.g., Blanco et al., 1990; Nagar et al., 2002; Ramos Almeida et al., 2008; Reunanen et al., 2003). Against the fact that the observations generally support the unified model, there is the question why some, but not all, type-2 AGN do not show any observational signs of PBLs. Tran (2001, 2003) found that only 30–50% of type-2 AGN show PBLs. Some have advocated that the non-detection of a PBL is due to genuine lack of a BLR (e.g., Tran et al., 2011). Others have suggested that the non-detection is due to obscuration effects, rendering the detection of PBLs as difficult or impossible, even with deep near-IR (NIR) spectro-polarimetry (Alexander, 2001). Using a statistically complete IRAS 60 µm selected type-2 AGN catalog, Heisler et al. (1997) investigated the relationship between the detectability of PBLs and the torus inclination angle. They showed that only AGN with a low torus inclination angle have high detection rate of PBLs compared to those with high inclinations. This result strongly suggests that PBLs could be obscured when there is an edge-on view through the torus and/or nuclear obscuration in the host galaxies. In addition to the optical spectro-polarimetry, X-ray observations suggest that there is a weak evidence showing different absorption in two types of

11 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

type-2 AGN. Gu et al. (2001) found that the AGN with PBL have slightly lower column density

(NH) than those without PBL. Similarly, Lumsden et al. (2004) showed that the detection rate

of PBL decreases as a function of NH, suggesting the absorption eﬀect by dusty torus could play a role of the detectability of PBL in AGN. To understand the role of the obscuration by the torus in type-2 AGN and the detectability of PBLs, knowing the torus geometry/morphology and properties is crucial. This is extensively discussed in Chapter 3.

1.5.3 Luminosity dependence of AGN torus

The more interest of obscuring torus is now stepping into its origin and its luminosity dependence, and even the connections of broad line regions (BLRs). X-ray surveys of AGN have revealed that the cosmological evolution of AGN luminosity function is well described by the luminosity-dependent density evolutions called “down-sizing,” where the density of lower- luminosity AGN show the peak at lower redshift than that of high-luminosity ones (e.g., Ueda et al., 2003). This fact suggests that the growth history of SMBHs may fundamentally differ between the low-mass and high-mass SMBH-galaxy systems in the present day universe. It is highly urged for astronomers to reveal the links between SMBHs and the tori as neighborhoods of them. Therefore, it is quite interesting to investigate the nature of AGN and tori in a wider luminosity range. Indeed, observed AGN and torus properties strongly depend on the lumi- 42 nosities in many aspects. For Lbol ≥ 10 erg/s, Lawrence (1991) presented that the fraction of type-2 AGN has a decreasing function of AGN luminosity called “receding torus model”. This trend is in good agreement with the results derived by using obscured AGN fraction in the hard X-ray (Hasinger, 2008; Ueda et al., 2003), type-2 AGN fraction in optical (Simpson, 2005), the ratio of reprocessed emission of AGN power (Ichikawa et al., 2012b; Maiolino et al., 2007). On the torus origins, statistical compilations of Seyfert AGN observations by Schmitt (2001) suggest a torus scale height (H) – radius(R) relationship has H/R ∼ 1. This means the AGN torus has dynamically stable structure to maintain the scale height (R ∼ H). One possible origin of vertical motions to sustain the torus clouds can be explained by the outflowing clouds confined by magnetic field from the disk (Elitzur & Shlosman, 2006; Elitzur, 2008), where the idea naturally arises based on disk-wind scenario, which successfully explains the outflow of clouds in a hydrodynamical method (Emmering et al., 1992; Everett, 2005; Konigl & Kartje, 42 1994). They pointed out that AGN torus may disappear in the low luminosity (Lbol ≤ 10 erg/s) range and sometimes even broad line region clouds may also disappear at the luminosity 39 with Lbol ∼ 5 × 10 erg/s due to the low accretion rate into the disk (Elitzur & Ho, 2009). Observational results in agreement with this theoretical model are reported from several au-

12 1.6 Co-Evolution of SMBHs and Their Host galaxies

thors. From NIR imaging polarimetry, Lopez-Rodriguez et al. (2013) measured the magnetic ﬁeld strength in the torus and showed that it is strong enough to lift up dust clumps. Conduct- ing high spatial MIR observations to LLAGN, Mason et al. (2012, 2013) obtained a low ratio of silicate feature strength over hydrogen column density, which is an indicator of dust-to-gas ratio inside the torus, consistent with the idea that the tori of LLAGN are in a disappearance phase. Beckmann et al. (2009); Burlon et al. (2011) reported the same trend that the fraction 42 of obscured AGN has a peak around L15−55keV ≤ 10 erg/s below which it starts to decrease towards lower luminosity. Utilizing the broadband X-ray spectra of Suzaku, Kawamuro et al. (2013) revealed LLAGN actually have larger opening angle of tori calculated from their equivalent widths of Fe Kα line. Thus, dividing the sample into the categories whether low/high luminosity AGN are crucial to further understand the AGN geometrical torus structures, the size, and its origins. This issue will be discussed in Chapter 5.

1.6 Co-Evolution of SMBHs and Their Host galaxies

Figure 1.9: MBH–σ relation for AGN (blue circles) and quiescent galaxies (red stars). - The ﬁgure is taken from Woo et al. (2013).

The apparent ubiquity of supermassive black holes (SMBHs) in the spheroidal components of present-day galaxies, and the correlation between the masses of SMBHs and spheroidal components (e.g., Magorrian et al., 1998) indicates that AGN (SMBH-driven activity) and

13 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

star-formation (SF; plausibly the progenitors of spheroids) are physically connected. Therefore, as shown in Figure 1.9, investigating AGN enable us to access the link of the formation and evolution of SMBHs and their host galaxies. While AGN present a variety of observational characteristics, the unified model for AGN propose the ubiquitous presence of a dusty, obscuring, and geometrically thick “torus” around the central engine and that all AGN are fundamentally same (Antonucci et al. 1985 and Chapter 1.4). The AGN luminosity dependence of torus covering factor (“receding torus model”, discussed in Chapter 1.5) showed that there is a strong connection between SMBH activity (∼ 0.001 pc) and the torus (∼ 1 pc). Then, who delivers the mass to the dusty torus? The primary candidate is the nuclear SF with ∼100 pc scale. To disentangle torus and SF activity in the galactic center, it is essential to trace only SF or AGN activity. An effective tool to pick up only SF activity is polycyclic aromatic hydrocarbon (PAH) emission feature around 3.3 µm in the L band (3– 4 µm) spectra. In a SF galaxy, PAHs are excited, without destruction, by far-UV photons from stars, and so a SF galaxy usually shows a strong 3.3 µm PAH emission feature. However, PAHs are destroyed because of the strong X-ray radiation from the AGN, and so a pure AGN shows no PAH emission. Thus, by using PAH emission feature, one can pick up only SF activity, by removing the contamination from an AGN (e.g., Imanishi & Dudley, 2000). Obtaining the data set of Subaru/IRCS L band spectroscopy of nearby AGN, Imanishi et al. (2011a) studied the connection between torus luminosity (L band or N band luminosity) and the nuclear SF activity obtained from 3.3 µm PAH emission feature. As a result, the authors showed that torus and SF luminosities are highly correlated each other. This result suggests that nuclear SF activity (∼100 pc) prompts to feed the torus (∼1 pc) and finally to the SMBH (∼0.001 pc). However, these studies are highly biased due to the optically identified AGN. Considering most AGN are obscured due to the covering of the dusty torus, more complete AGN surveys are urged to understand the complete view of AGN and the connection to the host galaxies.

1.7 Obscured AGN Survey

Recent cosmic X-ray background studies suggests that the population of highly obscured AGN 24 −2 with Compton-thick level (NH > 10 cm ) is not negligible as shown in Figure 1.10 (e.g., Ueda et al., 2014, and references there in). Utilizing X-ray data of the Chandra Deep Field South, Brightman & Ueda (2012) reported that the fraction of “Compton-thick” AGN (those with 24 −2 absorption column densities of NH > 10 cm ) increases with redshift from z = 0 to z > 1. This fact implies that the level of dust-obscuration of AGN may be correlated to the star formation activity, which has a peak at z ∼ 2 (Hopkins & Beacom, 2006). Indeed, Goulding et

14 1.7 Obscured AGN Survey

Figure 1.10: Cosmic X-ray Background spectrum calculated by Ueda et al. (2014). - The cosmic X-ray background spectrum calculated from the AGN population synthesis model (upper solid curve, red) by Ueda et al. (2014) compared with the observed data by various missions (Ajello et al., 2008). Middle solid curve (black): the integrated spectrum of Compton thin AGN with log NH < 24. Lower solid curve (red): that of Compton thick AGN with log NH = 24 − 26. Long-dashed curve (black): that of AGN with log NH = 23 − 24. Short-dashed curve (black): that of AGN with log NH = 22 − 23 Dot-dashed curve (black): that of AGN with log NH < 22. The figure is taken from Ueda et al. (2014). al. (2012) gathered all available Spitzer spectra of local Compton-thick AGN and found strong starburst (SB) features in their average infrared spectrum, suggesting that SB activity and highly obscured tori are somehow coupled each other. Some authors reported the same trend for buried AGN with small opening angles of tori (“new type” AGN; Eguchi et al., 2009, 2011; Ueda et al., 2007; Winter et al., 2009a) detected in the Swift/BAT survey. Although hard X-rays 24.5 −2 are useful to search for obscured AGN, extremely heavily buried AGN with NH > 10 cm are difficult to be detected due to flux attenuation by repeated Comptonization even at energies above 10 keV (e.g., Brightman & Nandra, 2011; Ikeda et al., 2009). Candidates of galaxies that host such buried AGN are infrared galaxies, defined as those 10 having large infrared luminosities (LIR) with LIR ≥ 10 L . Among infrared galaxies, those 11 12 in the range 10 L ≤ LIR < 10 L are called luminous infrared galaxies (LIRGs; Sanders 12 13 & Mirabel, 1996) and more luminous ones in the range 10 L ≤ LIR < 10 L are called ultraluminous infrared galaxies (ULIRGs; Sanders et al., 1988). Their bolometric luminosities are dominated by the infrared emission, suggesting that very luminous heating sources are surrounded by dust and then the heated dust re-emits in the infrared band. The hidden energy sources are believed to be either SB, AGN, or both. Disentangling the energy sources of infrared galaxies is crucial to unveil the history of dust obscured star formation and SMBH growth in

15 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

the universe (Goto et al., 2010; Le Floc’h et al., 2005; Magnelli et al., 2011; Murphy et al., 2011). Former studies Veilleux et al. (1995, 1999); Yuan et al. (2010) studied AGN activity in U/LIRGs on the basis of optical observations. They showed that the fraction of AGN increases 10 13 with infrared luminosities in the range of 10 L ≤ LIR < 10 L . One caveat of their optical diagnostics is that one cannot identify buried AGN embedded in dust-tori with very small opening angles where no significant narrow line regions can be formed. Unlike the AGN which are optically-classified type-2 Seyfert, such buried AGN are elusive to conventional optical spectroscopy (Maiolino et al., 2003). After the launch of Spitzer, the AGN activity in local U/LIRGs has been also addressed in the infrared band. Nardini et al. (2008, 2009, 2010) observed ULIRGs with Spitzer/IRS 5–8 µm spectroscopy, and investigated both the SB and AGN contributions by fitting SB/AGN templates to the spectra. They found the increasing AGN significance with infrared luminosity, confirming the same trend reported from the optical studies. Alonso-Herrero et al. (2012a) observed the 5–38 µm spectra of LIRGs with Spitzer. They decomposed the spectra into SB and AGN components using a SB galaxy template and clumpy torus models (Asensio Ramos & Ramos Almeida, 2009; Nenkova et al., 2008a,b), which are supported by observations (e.g., Gandhi et al., 2009; Ichikawa et al., 2012a). They confirmed that the AGN energetic contribution to the total power increases with infrared luminosity from 11 12 10 L ≤ LIR ≤ 10 L . The more discussion on buried AGN survey with AKARI/IRC 2.5–5.0 µm spectra will be discussed in Chapter 4.

1.8 Motivation and Strategy of This Study

As mentioned above, in order to fully understand the dusty torus properties and the connection between SMBH and the host galaxies, the complete AGN survey of hard X-ray and infrared are essential. To achieve this goal, in Chapter 2, the author ﬁrst works on the infrared/Hard X-ary catalog making of complete AGN survey in order to constrain the torus models on the market (smooth versus clumpy torus models). After the constraints of the torus models, in Chapter 3, the author applied the torus models to the observed SEDs of AGN torus using high spatial infrared observations obtained by 8-m class telescopes. Thanks to the combination of the torus models and the high quality data set of the torus SEDs, the author investigated how the torus geometry among AGN types aﬀects observed properties including the BLR detectability. Finally, in Chapter 4, the author found highly obscured (buried) AGN using AKARI /IRC 2.5–5.0 µm spectroscopies by decomposing the continuum into stellar and AGN dust-torus components. The author summarizes the conclusion of this dissertation and compiles the future

16 1.8 Motivation and Strategy of This Study

outlook based on this dissertation.

17 1. GENERAL INTRODUCTION ON ACTIVE GALACTIC NUCLEI

18 2

Mid- and Far-infrared Properties of a Complete Sample of Local Active Galactic Nuclei

2.1 Introduction

A complete survey of Active Galactic Nuclei (AGN) throughout the history of the universe is one of main goals in modern astronomy, which is necessary to understand the evolution of supermassive black holes (SMBHs) in galactic centers and their host galaxies. Given the fact that the majority of AGN are obscured by dust and gas surrounding the SMBH, observations in hard X-ray and mid-infrared (MIR) bands are proposed to be promising tools for detecting the whole populations of AGN (both radio quiet and loud ones) thanks to their strong penetrating power than optical/UV lights and soft X-rays. In fact, recent deep multi-wavelengths surveys utilizing these energy bands are discovering a large number of obscured AGN (Brandt & Hasinger, 2005; Brightman & Ueda, 2012; Brightman et al., 2014). Hard X-ray selection gives the most efficient way to have a clean AGN sample with little contamination from host galaxies. On the other hand, MIR selection sometimes achieves even better sensitivities in detecting AGN candidates than currently available X-ray data below 10 keV for heavily Compton thick AGN with column 24 −2 densities of NH > 10 cm , although separation of AGN components from star forming activities could always become an issue (Oyabu et al., 2011). To understand the efficiency and completeness of these surveys at different wavelengths, it is quite important to establish the relation between hard X-rays and infrared emission of AGN based on a large sample of nearby, bright AGN for which detailed studies can be made. The Swift/Burst Alert Telescope (BAT) survey (Tueller et al., 2008) is one of the most sensitive all sky surveys in the hard X-ray band (>10 keV), providing us with the least biased

19 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

sample of AGN in the local universe including heavily obscured ones, along with those by INTEGRAL (Bird et al., 2010; Winkler et al., 2003). Suzaku follow-up observations of BAT detected AGN have discovered a new type of deeply buried AGN with a very small scattering fraction (Eguchi et al., 2009; Ueda et al., 2007; Winter et al., 2009a). Assuming that the amount of gas responsible for scattering is not much diﬀerent from other objects, it is suggested that these new type AGN are obscured in a geometrically thick torus with a small opening angle. Understanding the nature of this population is important to reveal their roles in the cosmological evolution of SMBHs and host galaxies.

The MIR band also provides crucial information on the AGN tori. It is known that a thermal continuum in the MIR band originates from hot circumnuclear dust heated by optical/UV/X- ray photons from the central engine. Many works suggest that MIR emission is a good indicator of AGN activity. Many authors (Asmus et al., 2011; Gandhi et al., 2009; Horst et al., 2008; Krabbe et al., 2001; Levenson et al., 2009) have found a strong correlation between X-ray (2–10 keV) and MIR (12.3 µm) luminosity from the nucleus of Seyfert galaxies, using the 8-m class telescopes including VLT/VISIR and Gemini/T-ReCS data where the AGN torus emission can be spatially resolved from the host galaxy in many cases. In this Chapter, we statistically examine the correlation between the infrared and hard X-ray luminosities of AGN by utilizing a large uniform sample in the local universe, and investigate their infrared properties as a function of obscuration type. For this purpose, we use the Swift/BAT 9-month catalog (Tueller et al., 2008) as the parent sample, whose multiwavelength properties have been intensively investigated. Here we focus only on “non-blazar” AGN. As for the infrared data, we primarily use the all-sky survey catalogs obtained with AKARI, Japanese first infrared astronomical satellite launched on 2006 February 22 (Murakami et al., 2007), which provide unbiased galaxy samples selected in the mid- and far-infrared (FIR) bands with unprecedented sensitivities as an all sky survey mission. To complement the infrared data of AGN whose counterparts are not detected or do not have reliable flux measurements with AKARI, we also utilize the all sky survey catalog of NASA’s Wide-field Infrared Survey Explorer (WISE; Wright et al., 2010) mission, launched in 2010, as well as the catalogs of the Infrared Astronomical Satellite (IRAS; Neugebauer et al., 1984), a joint project of the US, UK, and the Netherlands launched on January 25, 1983. In Chapter 2.2, we present the sample selection criteria and the results of cross correlation between the Swift/BAT and AKARI catalogs. Throughout this Chapter, we −1 −1 adopt H0 = 70.0 km s Mpc ,ΩM = 0.3, and ΩΛ = 0.7.

20 2.2 Sample

2.2 Sample 2.2.1 Swift/BAT Hard X-ray Catalog

The Swift/BAT 9-month catalog (Tueller et al., 2008) contains 137 non-blazar AGN with a ﬂux limit of 2 × 10−11 erg cm−2 s−1 in the 14–195 keV band. The redshift range of this sample is 0 < z < 0.16. Winter et al. (2009a) investigated the soft X-ray (0.5–10 keV) properties of 128 (94.8%) BAT-detected non-blazar AGN of Tueller et al. (2008). By ﬁtting the X-ray spectra taken with Swift/X-Ray Telescope (XRT) or XMM-Newton, they derive key spectral parameters, such as the absorption column density (NH), covering fraction of the absorber (fc) or the scattering fraction (fs) with respect to the transmitted component (fscat ' 1 − fc). In our sample, we do not use the interacting galaxies NGC 6921 and MCG +04-48-002, which are not separated in the Swift/BAT catalog (e.g., Koss et al., 2011). Hence, the parent sample consists of 135 sources.

2.2.2 Infrared Catalogs 2.2.2.1 AKARI Point Source Catalogs

To obtain the infrared band properties of these Swift/BAT AGN, we mainly use the AKARI All-Sky Survey Point Source Catalogs (AKARI-PSC). AKARI carries two instruments, the infrared camera (IRC; Onaka et al., 2007) for the 2–26 µm band (centered at 9 µm and 18 µm) and the Far-Infrared Surveyor (FIS; Kawada et al., 2007) for the 50–200 µm band (centered at 65, 90, 140, and 160 µm). One of the major objective of AKARI satellite is to obtain an all-sky map of infrared sources. The AKARI all-sky survey observations covered nearly the full sky (≥96 %), and detected 870,973 sources with the IRC and 427,071 sources with the FIS. It achieved the flux sensitivities of 0.05, 0.09, 2.4, 0.55, 1.4, and 6.3 Jy with position accuracies of 6 arcsec at the 9, 18, 65, 90, 140, and 160 µm bands, respectively. In our study, we only utilize sources with the quality flag of fqual = 3, whose flux measurements are reliable1. As for the FIS catalog, we only refer to the 90 µm data as a representative FIR flux, which achieve the most significant sensitivity improvement compared with the previous IRAS mission among the four FIR bands.

2.2.2.2 IRAS Catalogs

The IRAS mission performed an unbiased all sky survey at the 12, 25, 60 and 100 µm bands. The typical position accuracy at 12 and 25 µm is 7 arcsec and 35 arcsec in the scan and cross

1 See the release note of the AKARI/FIS catalog for the details of fqual. It is recommended not to use the ﬂux data when fqual ≤ 2 for a reliable scientiﬁc analysis. http://irsa.ipac.caltech.edu/data/AKARI/documentation/AKARI-FIS BSC V1 RN.pdf

21 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

scan direction, respectively (Beichman et al., 1988). In this paper we use two largest catalogs, the IRAS Point Source Catalog (IRAS-PSC) and the IRAS Faint Source Catalog (IRAS-FSC). IRAS achieved 10σ point source sensitivities better than 0.7 Jy over the whole sky. The IRAS- FSC contains even fainter sources with ﬂuxes of >0.2 Jy in the 12 and 25 µm bands. We use only IRAS sources with fqual = 3 (the highest quality) 1.

2.2.2.3 WISE All-Sky Catalog

The WISE mission mapped the sky at the 3.4, 4.6, 12, and 22 µm bands, achieving 5σ point source sensitivity better than 0.08, 0.11, 1, and 6 mJy, respectively, in unconfused regions on the ecliptic poles (Wright et al., 2010). The WISE all-sky survey utilizes the data taken from 2010 January 7 to August 6 2. The source catalog contains positional and photometric information for over 563 million objects. The position accuracy estimated from the comparison with the 2MASS catalog is ∼2 arcsec at 3σ level. In this paper, we only use sources with the flux quality indicator ph qual = A, which has Signal-to-Noise ratio larger than 10. Since the angular resolutions of WISE (6.5 and 12.0 arcsec at 12 and 22 µm, respectively) are slightly worse than those of AKARI IRC (5.5 and 5.7 arcsec at 9 and 18 µm, respectively), we refer to the profile-fitting photometry of WISE derived by assuming point-like sources, for consistency with the AKARI catalog. This is justified because our targets of WISE are relatively distant (hence more compact) compared with those detected with AKARI or IRAS (see Chapter 2.2.3 for details). The photometric data of WISE are given in Vega magnitude, from which we convert into the Jansky unit using the zero-point flux densities of Fν (iso) = 31.674 Jy and 8.363 Jy for 12 µm and 22 µm, respectively.

2.2.3 Cross Correlation of Swift/BAT AGN with the IR Catalogs

We determine the IR counterparts of the Swift/BAT AGN by cross-correlating the AKARI, IRAS, and WISE catalogs in this order. Our primary goal is to obtain the photometric data in the MIR band as completely as possible from the hard X-ray selected sample. We put the highest priority to the AKARI catalog because of its high sky coverage (97 % of the all sky) and 2–4 times higher sensitivity than the IRAS survey. While all the IRAS sources should be detected with AKARI, AKARI ’s flux quality flags of very nearby (z < 0.005) objects turn out to be bad due to their extended morphology when fitted with a single Gaussian. In such cases, we rather refer to the IRAS data with good flux quality, which have ≈11 times worse angular

1 see Beichman et al. (1988) for the deﬁnition of fqual in the IRAS catalogs. False detections may be included when fqual ≤ 2. 2see the release note of WISE, http://wise2.ipac.caltech.edu/docs/release/allsky/

22 2.2 Sample

resolution than AKARI, since we aim to measure the total MIR ﬂux from both nucleus and host galaxy in a uniform way for all the AGN sample. For AGN that are not detected with AKARI or IRAS, we utilize the WISE all-sky catalog, which has 50 times better sensitivity than AKARI and therefore we can search fainter sources than ever in the MIR all-sky view, although the bright source are saturated due to the high sensitivity.

30 IRC FIS 25

Number of sources 5

0 0 0.05 0.1 0.15 0.2 Distance [arcmin]

Figure 2.1: Positional diﬀerence between Swift/BAT and AKARI sources - The histograms of position diﬀerence between the optical counterparts of the Swift/BAT AGN and their AKARI counterparts (red solid line: IRC, blue dashed line: FIS).

First, based on the positional matching of the optical counterparts of the Swift/BAT AGN with the AKARI-PSC, we determine their infrared counterparts in the 9 µm, 18 µm, and 90 µm bands. Here we adopt the maximum angular separation of 0.150 and 0.20 for the IRC and FIS sources, respectively, which correspond to typical 3σ positional errors at faintest fluxes (Ishihara et al., 2010; Yamamura et al., 2010). We find 70, 79, and 62 AKARI counterparts in the 9 µm, 18µm, and 90 µm bands out of the total 135 non-blazar BAT AGN sample. Figure 2.1 shows the distribution of the angular separation between AKARI and optical positions for the Swift/BAT AGN with IRC counterparts (red) and those with FIS counterparts (blue). The IRC sources are more concentrated in a small distance range (with an average of h∆ri = 0.020) than the FIS sources (h∆ri = 0.080), as expected from the positional accuracy in these catalogs. Further, for AGN whose MIR fluxes are not reliably measured (fqual < 3) or not detected with AKARI (65 and 56 sources in the 9 µm and 18 µm), we search for their counterparts at 12 µm or 25 µm in the IRAS-FSC and IRAS-PSC. Here we adopt the 50 arcsec radius, corresponding to the <2σ positional error in the cross-scan direction. As a result, 11 and 9 IRAS counterparts with fqual = 3 are identified in the 12 µm and 25 µm band, respectively.

23 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

Finally, we utilize the the WISE catalog to find the MIR counterparts in the 12 µm or 22 µm band for the remaining AGN detected neither in the AKARI nor IRAS catalogs (54 and 47 sources in the 9 µm and 18 µm bands). The matching radius of 2 arcsec is adopted. Thus, we identify 45 and 39 WISE sources with ph qual = A in the 12 µm and 22 µm band, respectively. In summary, we identify total 128 MIR counterparts detected any in the 9, 12, 18, 22, and 25 µm out of the 135 Swift/BAT AGN. Thus, the completeness of identification in the MIR band is 95%. We confirm that the probability of wrong identification with unassociated IR sources is negligible with these criteria for all the IR catalogs. Since the mean number density of the AKARI-PSC IRC and FIS sources in the all sky is ∼20 deg−2 and ∼10 deg−2, the expected number of contamination for the total 135 AGN within each error circle is estimated to be only 0.05 and 0.04, respectively. The number density in the IRAS catalog is 6.2 deg−2, and hence the expected contamination for the 65 AGN whose counterparts are searched for within the radius of 50 arcsec is 0.24. Similarly, we estimate false identification of the WISE sources with a number density of 5.63 × 108/(4.125 × 104) = 1.36 × 104 deg−2 to be 0.71 within the radius of 2 arcsec for the searched 54 sources.

2.2.4 AGN Type

To examine the infrared properties for different AGN populations, we divide the sample into three types based on the X-ray spectra. The first one is “X-ray type-1” (hereafter type-1) 22 −2 AGN, defined as those showing the absorption column density of NH < 10 cm . The second 22 −2 is “X-ray type-2” (hereafter type-2) AGN that have NH > 10 cm . In addition, we are interested in whether or not there is distinction in the IR properties of “new type” AGN, which

exhibit extremely small scattered fraction (fscat ≡ 1 − fc) suggesting the geometrically thick

tori around the nuclei. Here we deﬁne new type AGN as those satisfying fc ≥ 0.995 (Ueda et al., 2007), which are treated separately from the other (normal) type-2 AGN in this study.

2.2.5 Luminosity Correlation between AKARI, IRAS, and WISE Data

To make it possible to uniformly treat the MIR luminosities of AGN at slightly diﬀerent wavelengths obtained from the three IR observatories, we here investigate the correlation between the AKARI /IRAS/WISE luminosities, using the sample commonly detected with AKARI and IRAS, or with AKARI and WISE. We choose IRAS 12 µm/WISE 12 µm for AKARI 9 µm, and IRAS 25 µm/WISE 22 µm for AKARI 18 µm, respectively, because of the proximity of the central wavelengths. Figure 2.2 displays luminosity correlations between (1) AKARI 9 µm

24 2.2 Sample ] ] 45.5 45.5 ] 45.5 -1 -1 -1

45 45 45

44.5 44.5 44.5 m) [erg s m) [erg s m) [erg s µ µ 44 44 µ 44

43.5 43.5 43.5

43 43 43 (IRAS 12 (IRAS 25 (WISE 12 λ λ λ L L

42.5 L 42.5 42.5 λ λ λ

42 42 42 log log 42 42.5 43 43.5 44 44.5 45 45.5 log 42 42.5 43 43.5 44 44.5 45 45.5 42 42.5 43 43.5 44 44.5 45 45.5 -1 -1 -1 logλLλ(AKARI 9 µm) [erg s ] logλLλ(AKARI 9 µm) [erg s ] logλLλ(AKARI 18 µm) [erg s ]

] 45.5 -1

44.5 m) [erg s

µ 44

43.5

43 (WISE 22 λ

L 42.5 λ

42 log 42 42.5 43 43.5 44 44.5 45 45.5 -1 logλLλ(AKARI 18 µm) [erg s ]

Figure 2.2: Infrared Correlation plots - Correlation plots of infrared luminosities between AKARI 9 µm and IRAS 12 µm (top left, 53 sample), AKARI 9 µm and WISE 12 µm (top middle, 61 sample), AKARI 18 µm and IRAS 25 µm (top right, 56 sample), and AKARI 18 µm 22 −2 and WISE 22 µm (bottom, 70 sample). Squares (blue) represent type-1 AGN (NH < 10 cm ), 22 −2 circles (red) type-2 AGN (NH ≥ 10 cm ), and diamonds (green) new type AGN. The regression lines are given by eqs. (2.1), (2.2), (2.3), and (2.4) in Chapter 2.2.5.

versus IRAS 12 µm, (2) AKARI 9 µm versus WISE 12 µm, (3) AKARI 18 µm versus IRAS 25 µm, and (4) AKARI 18 µm versus WISE 22 µm. We check the strength of these luminosity correlations by using Spearman’s test. We obtain Spearman Rank coefficient (ρ) and null hypothesis probability P of (1) (ρ, P ) = (0.97, 2.37 × 10−34), (2) (ρ, P ) = (0.97, 3.6 × 10−37), (3) (ρ, P ) = (0.98, 5.7 × 10−40), and (4) (ρ, P ) = (0.99, ≤ 10−42). Since correlations between “luminosities” can be forced from those between “fluxes”, we also check the strength of the flux- flux correlations and obtain (1) (ρ, P ) = (0.93, 7.6 × 10−24), (2) (ρ, P ) = (0.92, 4.9 × 10−26), (3) (ρ, P ) = (0.96, 5.9 × 10−33), and (4) (ρ, P ) = (0.99, ≤ ×10−42). Thus, we confirm that the correlations in both luminosity and flux between different infrared catalogs are tight and significant. The standard deviation of the luminosity-ratio distribution between these two bands in the logarithm scale is found to be (1) 0.14 dex, (2) 0.17 dex, (3) 0.10 dex, and (4) 0.08 dex, respectively. The dispersion does not affect our conclusion on the MIR and hard X-ray luminosity correlation (Chapter 2.3.1). Based on the correlation, we derive the empirical formula to convert the IRAS or WISE luminosities at 12 µm, 22 µm, or 25 µm into the equivalent AKARI

25 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

luminosities at 9 µm or 18 µm as follows;

log λLλ(AKARI 9 µm) = log λLλ(IRAS 12 µm) − 0.051 (2.1)

log λLλ(AKARI 9 µm) = log λLλ(WISE 12 µm) + 0.057 (2.2)

log λLλ(AKARI 18 µm) = log λLλ(IRAS 25 µm) − 0.058 (2.3)

log λLλ(AKARI 18 µm) = log λLλ(WISE 22 µm) − 0.016 (2.4)

Assuming that AGN detected not with AKARI but with IRAS or WISE should follow the same correlations as examined here, we apply these conversion factors to derive the 9 or 18 µm “AKARI equivalent” luminosities for them so that we can discuss the correlation with hard X-rays in a uniform way regardless of the matched catalogs. Among the 128 Swift/BAT AGN with MIR counterparts, 126 and 127 objects have the ﬂux measurement in the 9 µm and 18 µm band, respectively. The 9 µm sample consists of 70 AKARI sources, 11 IRAS sources, and 45 WISE sources (126 sources in total), while the 18 µm sample consists of 79 AKARI sources, 9 IRAS sources, and 39 WISE sources (127 sources in total). The summed sample detected either in the 9 µm, 18 µm, or AKARI 90 µm band consists of 128 sources (9 µm; 126, 18 µm; 127, 90 µm; 62).

2.2.6 Basic Properties of the Sample

20 Number of Sources 15

0 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 redshift

Figure 2.3: Redshift Distribution - Redshift distribution of the whole non-blazer AGN in the Swift/BAT 9-month catalog (red solid line: 135 objects) and of those with the MIR counterparts (blue dashed line: 128 objects).

Figure 2.3 displays the redshift distribution of the 128 AGN with MIR counterparts (blue)

26 2.2 Sample

together with that of the whole (Tueller et al., 2008) sample of 135 AGN (red). Since our sample is highly complete (95%), we regard our current sample with MIR counterparts as representative of the whole population of hard X-ray selected AGN, and hereafter ignore any issues related to the incompleteness. Table 2.1 summarizes the infrared to X-ray properties of all the 135 Swift/BAT 9 month non-blazar AGN in Tueller et al. (2008), including 128 objects with MIR counterparts: (1) source No. in Tueller et al. (2008), (2) object name, (3) redshift, (4)–(6) infrared ﬂuxes (Fν ) at 9 µm, 18 µm, and 90 µm, (7)–(9) infrared luminosities (λLλ) at 9 µm, 18 µm, and 90 µm, (10) reference catalog for the IR data for 9 µm, 18 µm, and 90 µm, (11) hard

X-ray ﬂux in the 14–195 keV band, (12) hard X-ray luminosity in the 14–195 keV band (LHX),

(13) X-ray absorption column density (NH), (14) covering fraction in the X-ray spectrum (fc), and (15) the reference for the X-ray spectra. For AGN whose AKARI MIR fluxes are not available, we convert the infrared fluxes and luminosities with IRAS or WISE into those at 9 µm or 18 µm according to the formula given in Chapter 2.2.5. Columns (1), (2), (3), (11) are taken from Tueller et al. (2008)1 . The X-ray spectral information (columns 13–14) is basically adopted from Winter et al. (2009a), while we refer to the results obtained with Suzaku (Awaki et al., 2008; Bianchi et al., 2009; Eguchi et al., 2009, 2011; González-Mart´ın et al., 2011; Itoh et al., 2008; Risaliti et al., 2009; Tazaki et al., 2011; Turner et al., 2009; Winter & Mushotzky, 2010; Winter et al., 2009b) and those with XMM-Newton (Ballantyne, 2005; Noguchi et al., 2010) whenever available. When the information is not available in Winter et al. (2009a), we refer to Tueller et al. (2008). All luminosities in this table are calculated by using the redshift given in column (3). There are total 13 new type AGN out of the 135 AGN, for which asterisks are attached to the source No. in Table 2.1 (column 1). Figure 2.4 plots the hard X-ray luminosity distribution of our AGN sample with the MIR counterparts. Those for type-1 (dashed blue), type-2 (dotted red), and new-type (dot-dashed green) are separately plotted. Previous studies on Swift/BAT selected AGN (Burlon et al., 2011; Winter et al., 2009a) have shown that the X-ray luminosity distribution of type-1 AGN has a higher peak luminosity than that of type-2 ones, as expected from the well-known correlation that the fraction of absorbed AGN decreases against luminosity (e.g., Ueda et al., 2003). Although such trend may not be clear in Figure 2.4 due to the smaller sample size and coarse bin width (0.5 dex), the averaged logarithmic luminosity of the type-1 and type-2 AGN in our sample is 43.9 and 43.6, respectively, and a K-S test applied to their distributions returns a null probability of 4.0 × 10−2. Thus, albeit marginal, type-1 AGN are more luminous on average than type-2 AGN in our sample.

1The redshifts of two AGN, No. 13: 2MASX J02162987+5126246 and No. 53: 2MASX J06403799–4321211, are adopted from http://heasarc.gsfc.nasa.gov/docs/swift/results/bs58mon/

27 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

Number of Sources 25

0 40.5 41 41.5 42 42.5 43 43.5 44 44.5 45 45.5 Logarithm of X•ray Luminosity

Figure 2.4: Redshift Distribution - Distribution of the hard X-ray (14–195 keV) luminosity of our sample with MIR counterparts (total: black). The dashed blue, dotted red, and dot-dashed green ones correspond to those of the type-1, type-2, and new type AGN, respectively.

28 2.2 Sample (a) (a) (a) (a) (a) (a) (a) (g) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (b) (b) (b) (b) (b) (b) (b) (h) X-ray c f ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· H N HX L log HX f (I, I, X) 2.3 42.88 15.9 0.9970 (a) (I, I, X) 5.1 43.26 26.6 0.9130 (a) (I, I, X) 5.2 44.14 2.57 0.9660 (a) (I, I, X) 3.6 43.70 0.330 (I, A, X) 2.9 44.90 204 (A, A, X)(A, A, X) 4.7 5.2 44.39 43.63 0.023 1.50 (X, X, X) 3.9 44.35 0.030 (X, X, X) 3.6 43.84 1.74 (A, A, X) 7.8 44.17 63.1 0.9974 (e) (X, A, X) 2.3 44.84 0.031 (X, X, X) 5.6 44.03 0.001 (A, A, X) 5.3 44.11 0.020 (W, A, X) 3.2 43.56 28.2 0.9975 (a) (W, A, X) 5.9 43.39 46.9 0.9930 (a) (W, W, X) 3.7 43.26 0.056 (W, W, X) 3.5 44.17 0 (W, W, X) 3.7 43.77 0.027 (W, W, X)(W, W, X) 3.2 3.9 44.43 44.28 0.027 0.100 (W, W, X) 3.5 44.85 4.10 0.9670 (a) (W, W, X)(W, W, X) 2.9 2.4 43.95 44.52 85.0 3.27 0.9797 (e) (W, W, X)(W, W, X) 3.4 12.5 44.83 44.84 28.2 0.620 0.9730 (a) (W, W, X) 4.5 43.41 0.004 (W, W, X) 6.1 44.24 29.3 0.9914 (e) (W, W, X) 2.2 43.80 0.060 (W, W, X) 2.8 44.03 16.2 /BAT 9-month AGN catalog (90) IR Swift λ ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· λL (18) λ 44.88 λL (9) λ 43.60 43.74 42.2043.19 42.02 44.59 43.16 43.58 44.58 43.60 43.2443.81 43.33 43.69 43.52 43.69 43.7043.69 43.77 43.63 44.2243.95 44.08 43.88 43.0943.85 43.18 43.95 43.6944.07 43.98 44.01 44.70 44.68 42.76 42.63 43.65 43.68 44.3043.25 44.00 43.73 43.33 43.72 44.50 44.29 43.20 43.27 43.73 44.02 λL 90 f ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ······ ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· 18 f 0.152 9 ·················· ·················· ··· ·················· Infrared and X-ray Properties of the AGN in the z f Table 2.1: No.(1) Object 1** (2)2** NGC 235A3 Mrk 3484**5 Mrk NGC 352 4546 Fairall 9 NGC 526A 0.022229 0.136 0.015034 (3) 0.295 0.308 0.012125 0.014864 0.593 0.213 0.012 0.736 0.019097 0.04702 0.417 (4) 0.016 0.141 43.54 0.229 0.292 (5) 0.440 43.69 (6) 43.09 (7) (I, A, A) (8) 9.5 43.68 (9) 16.0 0.9960 (10) (c) (11) (12) (13) (14) (15) 78**9 NGC ESO 612 297-G01810 NGC 788 Mrk 1018 0.025201 0.081 0.029771 0.132 0.135 0.672 0.013603 0.04244 0.153 43.49 0.162 2.604 0.059 0.312 43.41 43.96 0.085 43.52 43.51 (W, 44.25 W, A) 4.9 (A, I, A) 43.85 3.2 65.1 43.81 0.9972 111 (a) 0.9945 (a) 11 LEDA 138501 0.0492 1213 Mrk15 590 2MASX J02162987+5126246 NGC 0.0288 931 0.02638 0.016652 0.080 0.349 0.227 0.763 2.430 43.86 43.90 43.70 (A, A, A) 7.3 43.66 0.360 1617 NGC18 985 ESO19 416-G002 ESO 198-024 QSO B0241+622 0.059198 0.044 0.023 0.043 0.0455 0.052 0.300 0.039 0.165 0.635 0.064 0.368 0.576 1.291 44.66 44.37 44.68 44.42 43.94 44.27 (A, (A, A, A, A) A) 7.3 3.7 44.52 44.20 0.742 0.389 20** NGC21 1142 2MASX J03181899+6829322 0.0901 0.017 0.027 0.028847 0.265 0.380 22 NGC 1275 0.017559 0.442 1.988 6.928 44.01 44.36 44.21 (A, A, A) 11.5 43.90 0.151 2324 PKS25 0326–288 NGC 1365 ESO 548-G081 0.108 0.01448 0.005457 0.248 2.234 0.097 5.364 0.968 80.384 43.70 43.42 43.78 42.68 44.25 43.18 (A, A, (I, A) I, A) 7.2 3.3 42.68 43.19 450 0 0.7400 (f) 27 PGC 13946 0.036492 0.015 0.036 28 2MASX J03565655–4041453 0.0747 0.020 0.048 29 3C 105 0.089 0.009 0.035 30 3C 111.0 0.0485 0.081 0.135 31 1H 0419–577 0.104 0.081 0.106 32 3C 120 0.03301 0.203 0.497 1.468 44.23 44.31 44.08 (A, A, A) 11.2 44.45 0.160 3334 2MASX35 J04440903+2813003 MCG 0.01127 –01-13-025 1RXS J045205.0+493248 0.090 0.145 0.029 1.427 42.93 0.015894 0.039 42.75 0.055 43.13 (A, W, A) 7.6 43.33 3.39 0.9900 (a) 36 XSS J05054–2348 0.035043 0.060 0.124 37 4U 0517+17 0.017879 0.166 0.748 0.957 43.60 43.95 43.36 (A, A, A) 7.8 43.75 0 3839 Ark40 120 ESO 362-G018 PICTOR A 0.012642 0.224 0.032296 0.035058 0.570 0.252 0.073 0.253 0.135 4546 NGC 2110 MCG +08-11-011 0.020484 0.340 0.007789 1.283 0.300 2.377 0.566 44.03 4.594 44.30 43.13 43.10 43.87 43.31 (A, A, A) (A, A, 11.1 A) 44.02 25.6 0.250 43.54 2.84 0.9520 (a) 4748 EXO 055620–3820.2 IRAS 05589+2828 0.03387 0.532 0.033 0.691 0.201 0.454 0.955 44.23 44.28 43.90 (A, A, A) 5.6 44.15 0 49** ESO50 005-G00451 Mrk 3 ESO 121-IG028 0.006228 0.537 0.520 0.0403 8.501 0.016 43.18 0.013509 0.037 0.322 42.86 1.892 2.939 43.38 43.64 (A, A, 44.10 A) 4.2 43.60 42.56 (A, A, A) 115 10.1 0.9973 43.62 (e) 110 52 ESO 490-IG026 0.0248 0.173 0.706

29 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI (i) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (k) (b) (b) (b) (b) (b) X-ray c f ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· H N HX L log HX f (I, A, X) 9.1 43.21 0.031 (X, X, X) 5.3 44.30 0 (A, A, X) 4.1 44.00 0.039 (X, X, X) 3.3 44.45 1.81 0.8800 (a) (A, A, X) 19.4 42.18 500 (A, X, X) 4.7 43.83 69.0 0.9974 (a) (W, A, X) 4.6 43.60 11.4 (W, A, X) 2.3 43.42 30.0 0.9950 (a) (W, W, X) 2.8 44.40 16.1 (W, W, X) 5.5 44.46 16.0 (W, W, X) 3.4 44.03 0.060 (W, W, X) 3.0 43.69 33.0 0.9940 (a) (W, W, X) 3.1 45.31 11.1 0.9870 (a) (W, W, X) 5.4 44.19 1.78 (W, W, X) 3.6 43.95 85.7 0.9983 (a) (W, W, X) 2.2 44.80 0.270 (W, W, X) 2.5 44.34 0.004 (W, W, X) 3.7 42.54 0.010 (W, W, X) 3.9 43.99 0.035 (W, W, X) 2.5 43.88 10.8 0.9910 (a) (W, W, X) 2.6(W, W, X) 40.81 5.8 3.30 44.26 0.6780 3.25 (a) (W, W, X) 2.5 43.71 0.040 /BAT 9-month AGN catalog (90) IR Swift λ ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· λL (18) λ ······ λL (9) λ 43.88 43.92 43.75 43.80 43.07 43.24 43.11 43.18 44.5543.37 44.66 43.58 43.7943.74 43.85 43.62 43.64 43.71 44.47 44.70 43.97 43.99 42.10 42.07 43.70 43.65 43.62 43.66 43.43 43.29 39.94 40.25 43.63 43.67 43.14 43.23 42.9143.36 42.78 43.45 43.11 λL 90 f ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· 18 f ······ 9 ·················· ·················· Infrared and X-ray Properties of the AGN in the z f Table 2.1: No.(1) Object 53 (2) 2MASX J06403799–4321211 0.061 0.032 0.068 (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) 54 2MASX J06411806+3249313 0.047 0.042 0.089 5556 Mrk58 6 Mrk 79 IGR J07597–3842 0.04 0.01881 0.022189 0.180 0.276 0.522 0.611 0.920 1.358 43.68 44.01 43.84 44.05 43.39 43.70 (A, (A, A, A, A) A) 6.6 4.7 43.72 43.72 0.006 3.26 0.9090 (a) 6061 Mrk 18 2MASX J09043699+5536025 0.037 0.014 0.040 0.011088 0.106 0.252 2.010 42.99 42.86 43.26 (A, I, A) 3.1 42.93 18.2 0.9700 (a) 62 2MASX J09112999+4528060 0.026782 0.030 0.067 63 IRAS 09149–6206 0.0573 0.407 0.792 1.735 45.03 45.02 44.66 (A, A, A) 3.2 44.40 0.850 64 2MASX J09180027+0425066 0.156 0.021 0.050 65 MCG –01-24-012 0.019644 0.104 0.263 6667 MCG +04-22-042 Mrk 110 0.032349 0.078 0.178 0.03529 0.073 0.107 6869 NGC 2992 MCG –05-23-016 0.008486 0.384 0.007709 1.391 0.299 1.277 0.826 43.31 9.220 43.11 43.57 43.25 42.83 43.60 (A, A, A) (A, A, 21.9 A) 6.6 43.54 42.94 1.60 1.19 0.4760 (a) 7071 NGC72 3081 NGC73 3227 NGC74 3281 2MASX75** J10384520–4946531 LEDA Mrk 093974 0.06 417 0.007956 0.003859 0.167 0.010674 0.444 0.023923 0.699 0.415 1.128 0.096 2.364 1.509 10.596 0.256 0.032756 42.89 6.011 42.69 0.068 0.941 43.21 43.54 0.152 43.62 42.80 43.80 43.74 43.04 43.07 43.70 43.61 (A, A, (A, A) A, A) (A, A, (A, A) 8.8 A, 12.9 A) 7.3 43.09 42.63 3.4 43.27 98.0 1.74 43.65 0.9937 0.8520 86.3 4.38 0.9810 (d) (a) 0.9860 (a) (a) 77 NGC 3516 0.008836 0.262 0.651 1.317 43.17 43.27 42.87 (A, A, A) 10.6 43.27 0.353 78 RX J1127.2+1909 0.1055 0.040 0.130 7980 NGC 3783 SBS 1136+594 0.0601 0.00973 0.041 0.502 0.083 1.530 2.716 43.54 43.72 43.27 (A, A, A) 16.1 43.53 0.570 0.7220 (a) 81 UGC 06728 0.006518 0.051 0.091 82 2MASX J11454045–1827149 0.032949 0.078 0.131 83 CGCG 041–020 0.036045 0.054 0.112 84 IGR J12026–5349 0.027966 0.166 0.614 1.630 44.00 44.26 43.99 (A, A, A) 4.0 43.86 2.34 8586** NGC Ark87 4051 34788 NGC89 4102 NGC90 4138 NGC91 4151 Mrk92 766 NGC 4388 NGC 4395 0.002335 0.02244 0.346 0.002823 0.885 0.091 0.002962 1.082 4.557 0.003319 0.102 0.044 3.287 1.032 42.12 0.075 54.050 0.012929 0.008419 3.629 2.161 42.80 0.001064 0.220 0.462 42.23 4.594 0.013 41.34 0.859 1.589 42.98 42.91 0.052 10.349 3.312 42.24 41.29 43.38 43.43 43.16 43.50 (A, 43.61 A, 42.15 43.72 A) (A, 42.56 A, A) 4.6 (W, 43.73 W, A) 43.61 (A, 2.4 A, A) 41.74 2.1 (A, A, (A, A) A, 37.4 41.62 A) 0.029 41.61 25.3 42.96 2.3 210 8.00 43.60 5.32 42.93 0.9880 36.2 0.9590 0.525 (a) 0.9920 (a) (a) 9495** NGC ESO96 4507 506-G027 XSS J12389–1614 0.025024 0.036675 0.114 0.053 0.011802 0.207 0.510 0.109 0.613 1.163 43.73 4.370 43.72 43.69 43.78 43.46 43.65 (A, A, A) (A, A, 13.2 A) 19.3 44.28 43.78 84.1 0.9981 34.3 0.9710 (j) (a) 97 NGC 4593 0.009 0.344 0.569 98 WKK 1263 0.02443 0.091 0.170 0.769 43.51 43.58 43.54 (W, A, A) 2.8 43.58 0.060 100101 SBS 1301+540102** NGC NGC 4945 4992 0.02988 0.015 0.001878 0.025137 0.022 8.811 0.059 9.945 103 MCG –03-34-064 0.016541 0.453 1.873 4.634 43.96 44.28 43.97 (A, A, A) 4.7 43.46 40.7 0.9610 (a)

30 2.2 Sample (l) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (o) (a) (a) (a) (a) (a) (b) (b) (b) (n) (b) (m) X-ray c f ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· H N HX L log HX f (A, I, X) 8.1 44.81 0.013 (A, A, X) 74.8 42.74 5.50 (A, A, X) 7.0 43.41 2.50 (A, A, X) 4.4 43.97 0.013 (A, A, X) 5.1 44.20 0.219 (A, A, X) 4.5 44.51 58.0 0.9905 (a) (X, X, X) 10.9 44.05 0.912 (A, A, X) 10.1 44.88 0.120 (A, A, X) 3.3 44.50 27.6 0.9260 (a) (A, A, X)(A, A, X) 5.0 9.7 45.14 44.42 0.177 0.015 (X, A, X) 2.9 43.62 0.825 (A, A, X)(X, X, X)(A, A, X) 3.6 13.9 2.7 44.80 44.10 43.12 0.121 2.47 1.23 (W, A, X) 3.0 43.78 0.150 (W, W, X) 8.6 44.28 0.240 (W, W, X) 3.9 45.09 0.139 (W, W, X) 4.1 44.53 45.0 (W, W, X)(W, W, X) 3.3 10.8 44.73 45.03 23.0 0.280 0.9360 (a) /BAT 9-month AGN catalog (90) IR Swift λ ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· λL (18) λ 43.76 λL (9) λ 43.40 43.20 43.41 43.54 43.99 44.13 44.1143.65 44.27 43.70 44.2043.23 44.40 43.06 44.40 44.41 44.50 43.95 44.35 44.48 44.62 44.61 44.31 44.39 45.1344.35 44.90 44.35 44.62 44.48 43.53 43.55 44.0944.36 44.19 44.30 λL 90 f ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ······ ··· ··· ··· ··· 18 f 0.240 9 ·················· ····················· Infrared and X-ray Properties of the AGN in the z f Table 2.1: No.(1) Object 104 (2) Cen A 0.001825 (3) 10.190 13.150 (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)Notes.— Tables 2.1 summarizeincluding the 128 infrared with to (14) MIR X-ray counterparts: properties of all the 135 (15) Swift/BAT 9 month non-blazar AGN in Tueller et al. (2008), 105 MCG –06-30-015 0.007749 0.280 0.591 1.035 43.08 43.11 42.65 (A, A, A) 7.5 43.00 0.190 106107 NGC 5252 4U 1344-60 0.022975 0.012879 0.104 0.207 0.136 0.556 0.416 43.51 43.35 43.22 (W, W, A) 6.6 43.90 4.34 0.9620 (a) 108 IC 4329A 0.016054 0.769 1.790 1.785 44.16 44.23 43.53 (A, A, A) 30.0 44.24 0.610 109 Mrk 279 0.030451 0.141 0.387 110112 NGC 5506 NGC 5548 0.006181 0.01717 0.823 2.240 0.157 8.413 0.409 1.073 43.36 43.54 43.50 43.65 43.37 43.37 (A, A, A) (A, A, A) 23.6 5.8 43.30 43.59 2.78 0.070 0.9893 (a) 113 ESO 511-G030 0.02239 0.064 0.090 0.847 43.27 43.14 43.50 (W, W, A) 4.7 43.73 0.098 115 NGC 5728 0.0093 0.176 0.418 11.383 43.05 43.12 43.86 (A, A, A) 8.9 43.23 82.0 116117 Mrk 841 Mrk 290 0.036422 0.029577 0.126 0.085 0.372 0.151 118119 Mrk 1498 2MASX J16481523–3035037 0.031 0.030 0.038 0.0547 0.067 0.214 120 NGC 6240 0.02448 0.350 1.489 23.100 44.20 44.53 45.02 (A, A, A) 4.7 43.81 102 122 NGC 6300 0.003699 0.277 1.336 14.928 42.44 42.82 43.17 (A, A, A) 9.1 42.44 21.5 123 GRS 1734–292 0.0214 124 1RXS J174538.1+290823 0.111332 0.030 0.059 125126** 3C ESO 382 103-035127 3C 390.3 0.013286 0.300 1.446 0.05787 1.227 0.0561 0.120 43.59 0.106 0.090 0.242 43.97 43.20 (A, A, A) 9.7 43.58 21.6 0.9990 (a) 128129 NVSS J193013+341047130 NGC 6814131 3C 403 Cyg A 0.0629 0.130 0.254 0.005214 0.334 0.258 0.059 0.05607 6.954 0.093 0.152 42.65 0.215 0.418 42.41 2.455 44.47 43.14 44.72 (I, A, A) 44.79 6.2 (W, A, A) 42.57 10.9 0.058 44.91 11.0 133136 NGC 6860137 4C +74.26 Mrk 509 0.014884 0.104 0.155 0.357 0.0344 0.147 1.369 0.175 0.247 43.41 0.499 43.47 43.35 (A, A, A) 4.9 43.39 0.100 138139 IC 5063140 2MASX J21140128+8204483141 IGR 0.084 J21247+5058142 IGR J21277+5656144 RX J2135.9+4728 0.071 UGC 11871 0.105 0.02 0.0147 0.025 0.011348 0.211 1.159 0.435 2.246 0.026612 3.821 0.146 43.87 0.320 5.040 44.03 43.90 43.56 43.94 (I, A, A) 44.43 7.1 (A, A, A) 43.31 3.9 25.0 43.80 0.9910 2.44 (g) 145** NGC 7172146 NGC 7213 0.008683 0.316 0.005839 0.424 0.360 8.087 0.743 43.24 2.943 43.07 42.95 42.77 43.65 42.86 (A, A, A) (A, 12.4 I, A) 43.32 5.2 8.19 42.59 0.9990 0.025 (a) 147 NGC 7314 0.00476 0.268 0.304 4.499 42.48 42.41 42.88 (I, A, A) 5.7 42.46 1.16 148** NGC 7319149 3C 452 0.022507 0.088 0.158 0.0811 0.576 43.53 0.029 0.070 43.48 43.34 (A, A, A) 4.1 43.67 86.6 0.9965 (a) 151 MR 2251–178 0.06398 0.090 0.148 152153 NGC 7469 Mrk 926 0.016317 0.767 0.04686 2.692 0.060 27.694 44.18 0.214 0.647 44.42 44.01 44.74 44.26 (A, A, 44.04 A) 8.3 (A, A, A) 43.70 5.5 0.041 44.45 0.035 154 NGC 7582 0.005254 1.368 3.287 60.906 43.43 43.51 44.08 (A, A, A) 6.7 42.61 33.0

31 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

2.3 Results and Discussion

2.3.1 Correlation between the Infrared and Hard X-ray Luminosities

45.5 45.5 45.5

45 45 45 ] ] ]

-1 44.5 -1 44.5 -1 44.5

44 44 44

43.5 43.5 43.5 [erg s [erg s [erg s 43 43 43 HX HX HX

42.5 42.5 42.5

log L 42 log L 42 log L 42

41.5 41.5 41.5

41 41 41 41 41.5 42 42.5 43 43.5 44 44.5 45 45.5 41 41.5 42 42.5 43 43.5 44 44.5 45 45.5 41 41.5 42 42.5 43 43.5 44 44.5 45 45.5 -1 -1 -1 log λLλ(9 µm) [erg s ] logλLλ(18 µm) [erg s ] logλLλ(90 µm) [erg s ]

Figure 2.5: Mid-Infrared and Hard X-ray luminosity Correlations - Correlation between the infrared (9 µm, 18 µm, or 90 µm) and hard X-ray (14–195 keV) luminosities. Squares 22 −2 22 −2 (blue) represent type-1 AGN (NH < 10 cm ), circles (red) type-2 AGN (NH ≥ 10 cm ), and diamonds (green) new type AGN. Dotted lines represent the regression lines (see Section 2.3.1). The small-filled, large-filled, small-open symbols denote those with the AKARI, IRAS, and WISE counterparts, respectively. The IRAS and WISE fluxes are all converted into the 9 µm (left) or 18 µm (right) using the formula given in Section 2.5.

Figure 2.5 shows the luminosity correlations in the luminosity range from 1041 erg s−1 to 1046 erg s−1 between the infrared (9, 18, or 90 µm) and Swift/BAT hard X-ray bands. Type- 1, type-2, and new type AGN are marked with squares (blue), circles (red), and diamonds (green), respectively. The small-filled symbols denote the data from AKARI, large-filled ones those from IRAS, and small-open ones those from WISE. In the figure, NGC 4395 is not shown due to its low luminosities (log λLλ(9 µm), log λLλ(18 µm), log LHX) = (39.98, 40.28, 40.81). As noticed from the figure, the MIR (both 9 and 18 µm) luminosities well correlate with hard X-ray luminosity over 3 orders of magnitude (from 1042– 1045 erg s−1). In the FIR (90 µm) band, by contrast, the correlation is much weaker with larger dispersion compared with the MIR bands, even though we plot here only for AGN detected with AKARI at 90 µm. Least-square fits to the hard X-ray versus MIR luminosity plots with a power law model 43 (i.e., a linear function for the logarithmic luminosities with the form of log(LHX/10 ) = a + 43 b log(λLλ(9, 18 µm)/10 ) give the following best-fit correlations:

L λL (9 µm) log HX = (0.06 ± 0.07) + (1.12 ± 0.08) log λ (2.5) 1043 1043 L λL (18 µm) log HX = (0.02 ± 0.07) + (1.10 ± 0.07) log λ (2.6) 1043 1043

To check the signiﬁcance of the correlations between the hard X-ray and MIR/FIR luminosities (or ﬂuxes), we perform Spearman’s tests for the summed sample consisting of all AGN

32 2.3 Results and Discussion

types. The results are summarized in Table 2.2, which has the following columns: Col. (1) sample; Col. (2) number of objects; Col. (3) luminosity-luminosity correlation coeﬃcient (ρL);

Col. (4) flux-flux correlation coefficient (ρf ); Col. (5) a standard Student’s t-test null significance level for luminosity-luminosity correlations (PL); Col. (6) a standard Student’s t-test null significance level for flux-flux correlations (Pf ); Col. (7) regression intercept (a) and its 1-σ uncertainty; Col. (8) regression slope (b) and its 1-σ uncertainty. We find that both luminosity- luminosity and flux-flux correlations between the hard X-ray and MIR bands are highly significant. While there is also a significant correlation between 90 µm and hard X-ray luminosity, −7 (ρL,PL) = (0.59, 4.5 × 10 ), their flux-flux correlation is weak with (ρf ,Pf ) = (0.17, 0.18).

Table 2.2: Correlation parameters between MIR and the hard X-ray sample

Sample N ρL ρL PL Pf a b

(1) (2) (3) (4) (5) (6) (7) (8) 9 µm 126 0.82 0.60 3.0 × 10−31 1.4 × 10−13 0.06 ± 0.07 1.12 ± 0.08 18 µm 127 0.76 0.59 1.7 × 10−25 2.4 × 10−13 0.02 ± 0.07 1.10 ± 0.07 90 µm 62 0.59 0.17 4.5 × 10−7 1.8 × 10−1 −0.21 ± 0.10 1.16 ± 0.11 Notes.— Correlation properties between 14-195 keV X-ray luminosity and infrared luminosities to various subsample populations.

We establish the good correlation between the MIR and hard X-ray luminosities in AGN from so far the largest, uniform AGN sample in the local universe, although similar results have been reported by several authors (Gandhi et al., 2009; Mushotzky et al., 2008; Vasudevan et al., 2010). The MIR emission from galaxies hosting an AGN is believed to originate mainly from high temperature (∼ 150–300 K) dust emission heated by X-ray/UV photons from the central engine. Thus, if extinction is not important, the MIR luminosity is expected to be correlated with the intrinsic X-ray luminosity, the most reliable tracers of the AGN power (since our 24 −2 sample contains mostly Compton thin AGN with NH < 10 cm , the observed 14–195 keV luminosity can be regarded as the intrinsic one without any correction). On the other hand, the FIR emission comes both from cooler (∼ 30 K) interstellar dust heated by stars in host galaxies and from the cooler (outer) part of the torus in the AGN. The contribution from the host galaxy increases the scatter of the observed luminosity correlation, depending on the total star forming rate over the whole galaxy. Our results demonstrate that the MIR emission is more suitable for estimating the AGN intrinsic power than in the FIR band. It is interesting to check the consistency of our results with the luminosity dependence of obscured AGN fraction found in many previous works (e.g., Hasinger (2008); Maiolino et al.

(2007); Ricci et al. (2013); Simpson (2005); Ueda et al. (2003, 2014)). If LMIR totally originates from the reprocessed emission in the torus irradiated by an AGN, the average ratio of LMIR/Lbol 1.12 should represent the obscured fraction, or torus covering fraction. Our result LHX ∝ L (9 µm)

33 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

0.89 gives L(9 µm) ∝ LHX . Marinucci et al. (2012) derive a non-linear relation between 2–10 keV 1.18 1.18 and bolometric luminosities of AGN as Lbol ∝ L2−10 keV(∝ LHX ). Combining the above relations and assuming L2−10 keV ∝ LHX, we can estimate that the obscured AGN fraction is proportional to

0.89 L(9 µm) LHX −0.29 −0.25 ∝ 1.18 = LHX ∝ Lbol . (2.7) Lbol LHX This result suggests that the obscured fraction decreases with bolometric luminosity, conﬁrming the trend found by various AGN surveys.

30 18 18 (a) 16 (b) 16 (c) 25 14 14

20 12 12 Number of Sources Number of Sources Number of Sources 10 10 15 8 8

10 6 6

4 4 5 2 2

0 •1 •0.5 0 0.5 1 0 •1 •0.5 0 0.5 1 0 •1 •0.5 0 0.5 1 log(λL (9 µm)/L ) log(λL (9 µm)/L ) log(λL (9 µm)/L ) λ HX λ HX λ HX

22 20 (d) 12 (e) 12 (f) 18 10 10 16

14 8 8 Number of Sources Number of Sources Number of Sources 12 10 6 6 8 4 4 6

4 2 2 2

0 •1 •0.5 0 0.5 1 0 •1 •0.5 0 0.5 1 0 •1 •0.5 0 0.5 1 log(λ L (18 µm)/L ) log(λ L (18 µm)/L ) log(λ L (18 µm)/L ) λ HX λ HX λ HX

Figure 2.6: Histgram of mid-infrared to hard X-ray luminosity ratio - Histograms of log(λLλ(9 µm)/LHX) (top) and log(λLλ(18 µm)/LHX) (bottom). (a) and (d) (left): total AGN (solid lines), (b) and (e) (center): type-1 AGN, (c) and (f) (right): type-2 + new type AGN That of log(λLλ(12.3 µm)/LHX) for the Gandhi et al. (2009) sample is overplotted in panel (a) and (d) (dashed lines, hatched area).

Table 2.3: Average and standard deviation of log(λLλ(9, 18µm)/LHX)

Sample N r¯ σ (1) (2) (3) (4) 9 µm All 126 −0.129 ± 0.039 0.437 ± 0.055 Type-1 55 −0.165 ± 0.052 0.389 ± 0.074 Type-2(+ New Type) 71 −0.101 ± 0.056 0.470 ± 0.079 18 µm All 127 −0.080 ± 0.042 0.473 ± 0.059 Type-1 57 −0.141 ± 0.060 0.452 ± 0.085 Type-2(+ New Type) 70 −0.031 ± 0.058 0.483 ± 0.082 Averages and standard deviations of log(λLλ(9, 18 µm)/LHX).

Figures 2.6 plot the histograms of the MIR to hard X-ray luminosity ratio in the logarithm

34 2.3 Results and Discussion

scale (r ≡ log λLλ(9 µm)/LHX and log λLλ(18 µm)/LHX) for the Swift/BAT AGN sample with MIR counterparts in the 9 µm and 18 µm bands, respectively, calculated from (a, d) total, (b, e) type-1 AGN, and (c, f) type-2 + new type AGN. Following Gandhi et al. (2009), we calculate its average and standard deviation for each sample, which are summarized in Table 2.3. From the total AGN, we obtain (¯r, σ) = (−0.129 ± 0.039, 0.437 ± 0.055) for the 9 µm band and (¯r, σ) = (−0.080 ± 0.042, 0.473 ± 0.059) for the 18 µm band. We ﬁnd that the average is consistent between type-1 and type-2 (plus new type) AGN within the errors.

102 102

101 101 HX HX

0 m) / L 0 m) / L 10 10 µ µ (9 (18 λ λ L L

λ -1 -1 10 λ 10

10-2 10-2 10-3 10-2 10-1 100 101 102 103 10-3 10-2 10-1 100 101 102 103 22 -2 22 -2 NH [10 cm ] NH [10 cm ]

Figure 2.7: MIR to hard X-ray luminosity ratio as a function of NH - The MIR to hard X-ray luminosity ratio plotted against the absorption column density. Left: λLλ(9 µm)/LHX. Right:λLλ(18 µm)/LHX. All symbols are the same as Figure 2.5.

For comparison, the histogram of log λLλ(12.3 µm)/LHX obtained by Gandhi et al. (2009) from 42 local AGN are overplotted in Figures 2.6(a) and (d), where we convert their 2–10 keV luminosities into the 14–195 keV band by assuming a power law photon index of 2. Their sample is selected from nearby AGN with good available X-ray spectra observed with Suzaku, INTEGRAL, or Swift, whose average distance is 74 Mpc (or z = 0.017), consisting of 12 Seyfert 24 −2 1s, 19 Seyfert 2s, 3 LINERs, and 8 Compton thick (NH ≥ 1.5 × 10 cm ) AGN. It is not a statistically complete sample, however. The average X-ray luminosity is ∼ 42.9 and only 3 sources have QSO class luminosity (L(2 − 10 keV) > 1044erg s−1). Twenty four objects out of the 42 AGN are listed in the Swift/BAT 9-month catalog and in our Table 1. We obtain (¯r, σ) = (0.086 ± 0.054, 0.353 ± 0.077) for the 12.3 µm band from the Gandhi et al. (2009) sample. Although the average is slightly larger than ours, it is still consistent within the error in the 18 µm band. Note that silicate features (see below) may aﬀect our result in the 9 µm band. The standard deviation is almost the same between the Gandhi et al. (2009) sample and ours; the small diﬀerence could be due to the fact that there are 22 “well-resolved” AGN in Gandhi et

35 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

al. (2009), whose nucleus MIR emission was spatially separated from the host galaxy, making the correlation between the MIR and X-ray luminosities tighter. We also check the consistency with the results by Matsuta et al. (2012) obtained from the cross correlation between the Swift/BAT 22 month sample and the AKARI catalog. We conﬁrm

that our best-fit slopes (b) of the linear correlations between log λLλ(9, 18 µm) and log LHX are well consistent with their results (1.13 for 9 µm and 1.12 for 18 µm with average errors of 0.04). On the other hand, the averaged MIR to X-ray luminosity ratios (¯r) derived from our study are slightly smaller than those by Matsuta et al. (2012) (0.14 for 9 µm and 0.19 for 18 µm). The difference is attributable to the high completeness of identification in our analysis, where faint, WISE-only detected MIR sources are included. As noticed from Figure 2.5, all types of AGN (type-1, type-2, and new type) seem to follow almost the same MIR vs hard X-ray luminosity correlation. In fact, we see no significant difference in the distribution of their luminosity ratio between type-1 and type-2/new type

AGN (Table 3). To check this further, we plot λLλ(9 µm)/LHX and λLλ(18 µm)/LHX as a function of the absorption column density (NH) in Figure 2.7. In both panels, there is no clear dependence of the MIR to X-ray luminosity ratio on NH up to log NH ' 24. The large ratios found for Compton thick AGN can be partially explained by attenuation of the hard X-ray

fluxes due to heavy obscuration. The absence of NH dependence suggests that the emission from the AGN-heated dust seems not to be affected by the obscuration by the torus causing the X-ray absorption. The results cannot be explained with homogeneous dust torus models (Pier & Krolik, 1992, 1993), which predict the significant decrease in the MIR to X-ray luminosity ratio when an optically-thick line-of-sight through the torus primarily show cooler smooth-dust and a lower mid-infrared luminosity for the same X-ray luminosity than does an optically-thin one. Our results rather favor the clumpy dust tori model (Höniget al., 2006; Nenkova et al., 2008a; Schartmann et al., 2008) , which predicts isotropic MIR emission, as discussed in Gandhi et al. (2009). No clear difference is seen in the MIR to X-ray luminosity correlation between normal AGN and new type AGN. New type AGN may have large intrinsic MIR luminosities because of the geometrically thick torus (Nenkova et al., 2008b), which could be partially canceled out due to extinction, however. Currently, the possible contribution from the host galaxy in the MIR band makes direct comparison between individual objects very difficult. We need more observations of these AGN by resolving only the nucleus emission as done in Gandhi et al. (2009). Meléndez et al. (2008a,b); Weaver et al. (2010) proposed the [Ne v] 14.32/24.32 µm and [O iv] 25.89 µm lines as good AGN indicators because of their high ionization potential. Using a Swift/BAT sample, Weaver et al. (2010) found a tight luminosity correlation between these

36 2.3 Results and Discussion

lines and hard X-ray luminosities with a scatter of ∼0.5 dex, which is almost comparable with our result obtained for the MIR and hard X-ray correlation (≈0.45 dex, see Table 2.2). The correlation we ﬁnd is very useful and easy to apply to various survey data as it requires only photometry without spectroscopy.

2.3.2 Averaged IR Spectral Energy Distribution

Spectra in the MIR band are quite useful to investigate the host galaxy properties of AGN. It is known that Polycyclic Aromatic Hydrocarbon (PAH) molecule features in the 3 µm to 20 µm band (Tielens, 2008) can be used as a starburst tracer (Imanishi & Dudley, 2000). The AKARI 9 µm band covers the strong PAH emission feature at 7.7, 8.6, and 11.2 µm. In addition, opacity peak of amorphous silicate grains are located around 10 and 18 µm due to the Si−O stretching and the O−Si−Y bending modes. Hao et al. (2007) report the MIR spectra of different types of AGN (type-1 QSO, type-1 Seyfert, type-2 Seyfert, ULIRG) from Spitzer observations, revealing a large variety in the 10 µm silicate feature. Silicate absorption feature is clearly detected from type-2 Seyferts. Averaged SEDs in the NIR to FIR band of the three types of AGN are presented in Fig- ure 2.8. The spectrum of each AGN is normalized by the 18 µm luminosity. Here we only use total 42 sources detected in all the 9, 18, and 90 µm bands, consisting of 16 type-1, 21 type-2, and 5 new type AGN. We also include the photometric data in the J, H, and Ks bands adopted from the 2 Micron All Sky Survey (2MASS) Point Source Catalog. We neglect the effect of redshift because the sample consists of only local AGN (z < 0.1 with a mean value of hzi = 0.0165), for which K-correction is not significant. We find that the FIR emission at 90 µm is weaker relative to 18 µm in the type-1 AGN (blue) than in type-2 (red) and new type AGN (green). On the other hand, the MIR spectra are almost the same between the type-1 and type-2 AGN. This can be explained because type-1 AGN have intrinsically higher AGN luminosities on average than type-2 AGN (see Figure 2.4), and hence contribution from cool dust in the host galaxy emitting the FIR radiation becomes smaller relative to the AGN component mainly observed in the MIR band. Indeed, the same trend is seen in the observed SED templates of Seyfert 2 galaxies and type-1 QSOs complied by Polletta et al. (2007) (see their Figure 1). Another possible explanation would be an intrinsic difference of the dust quantity between type-1 and type-2 AGN: Malkan et al. (1998) suggested that the host galaxies of type-2 AGN are likely to be more dusty than those of type-1 AGN from the results of a Hubble Space Telescope imaging survey of nearby AGN. It is remarkable that the averaged SED of new type AGN exhibit enhanced fluxes than type-2 AGN at 9 µm. In fact, this is confirmed in the individual SED for 4 objects (ESO

37 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

1 Normalized Luminosity

0.1 1 10 100 Wavelength[µm]

Figure 2.8: Averaged IR SED - Averaged infrared (1–100 µm) SED normalized at the 18 µm band for type-1 (blue, ﬁlled square), type-2 (red, ﬁlled circle), and new type AGN (green, diamond).

005-G004, ESO 506-G027, NGC 7172 and NGC 7319) out of the 5 new type AGN examined here. This 9 µm excess most probably attributable to the PAH emission feature at 7.7, 8.6, and 11.2 µm from the host galaxies. This is consistent with the larger 90 µm excess in the averaged new-type AGN spectrum than in type-2 AGN, which also reﬂects the starburst activity in the host galaxies. Another possibility is that the 9 µm excess comes from the emission feature of silicate grains, which would imply a larger amount of dusts around the nucleus than in normal type-2 AGN. Our result suggests that geometrically thick tori around the black hole may form in galaxies with high star forming rates.

Four out of the 5 new type AGN have available MIR spectra observed with Spitzer. ESO 005-G004 shows a clear 11.2 µm PAH line, and silicate absorption features are suggested at λ > 10 µm (Weaver et al., 2010). A detection of the 11.2 µm PAH line is reported from ESO 506-G027 (Sargsyan et al., 2011), although the spectrum is not available in the literature. NGC 7172 shows PAH lines at 7.7 µm and 11.2 µm with strong silicate absorption features. In addition, for ESO 005-G004 and NGC 7172, the line ﬂux ratio between [Ne iii] 15.56 µm and [Ne ii] 12.81 µm is available in Weaver et al. (2010), from which the relative strength of the starburst to AGN activities can be estimated. These two sources have log f[NeIII]/f[NeII] = −0.50 and −0.31, respectively, suggesting that both have relatively strong starburst components normalized by their AGN activities. Thus, though limited in sample, a majority of our new type AGN indeed show signiﬁcant PAH emission lines (and silicate absorption features) contributing

38 2.4 Summary and Conclusions

to the 9 µm excess. Our results are also consistent with the study of Goulding et al. (2012), which suggest that the majority of highly obscured AGN show the star formation dominated MIR spectra. Further systematic investigation of the MIR spectra of hard X-ray selected AGN is useful to reveal the host galaxy properties and environment around the central engine in relation to the X-ray spectral information.

2.4 Summary and Conclusions

We have systematically studied the MIR and FIR properties of a large complete ﬂux limited AGN sample in the local universe detected in the Swift/BAT all sky survey in the 14–195 keV band, which has the least bias against obscuration. Utilizing the AKARI, IRAS, and WISE infrared catalogs, we unambiguously identify 128 counterparts in the MIR band out of the 135 non-blazar AGN in the Swift/BAT 9-month catalog by Tueller et al. (2008). For our discussion, the whole sample is divided into 3 types based on the X-ray spectra, 57 type-1, 58 type-2, and 13 “new type” AGN showing extremely small scattered fractions. The two main conclusions are summarized as follows:

1. We ﬁnd a good luminosity correlation between the MIR (9 µm and 18 µm) and hard X-ray

band over three orders of magnitude (42 < log LHX < 45), while that between the FIR (90 µm) and hard X-ray bands is weaker, most probably due to the larger contribution from the host galaxy (Figure 2.5). All types of AGN follow the same correlation, and the luminosity ratio between the MIR to X-ray bands show no clear dependence against 24 −2 absorption column density up to NH ∼ 10 cm . Our results favor isotropic infrared emission models, possibly clumpy dust torus models rather than homogeneous dust model, conﬁrming the argument by Gandhi et al. (2009) but with a much larger sample.

2. We ﬁnd 9 µm excess in the averaged infrared SED of “new type” AGN. This could be attributable to the PAH emission features, as conﬁrmed in the available Spitzer spectra of at least three sources, suggesting that their host galaxies have strong starburst activities.

39 2. MID- AND FAR-INFRARED PROPERTIES OF A COMPLETE SAMPLE OF LOCAL ACTIVE GALACTIC NUCLEI

40 3

Clumpy torus modeling of polarized and non-polarized BLR AGN

3.1 Introduction

While AGN present a variety of observational characteristics, the unified model for AGN proposes the ubiquitous presence of an obscuring torus around their central engines, and that all AGN are fundamentally the same (Antonucci, 1993). This optically and geometrically thick torus produces the effect of a line of sight viewing angle dependency. Type-1 AGN are observed with the direct view of fast moving material close to the supermassive black hole (SMBH), re- sulting in broad emission lines in their spectra, while type-2 AGN are observed from an edge-on view and the torus blocks the broad emission line region (BLR) component from our line of sight. The most compelling evidence of the unified model was the detection of polarized broad emission lines (PBLs) in type-2 AGN (e.g., Antonucci & Miller, 1985). Further evidence supporting the unified model comes from infrared (IR) observations in several type-2 AGN which showed the existence of obscured/hidden broad line regions (HBLRs) detectable only with dust penetrating infrared observations (e.g., Blanco et al., 1990; Nagar et al., 2002; Ramos Almeida et al., 2008; Reunanen et al., 2003). Against the fact that the observations generally support the unified model, there is the question why some, but not all, type-2 AGN do not show any observational signs of PBLs. Tran (2001, 2003) found that only 30–50% of type-2 AGN show PBLs. Some have advocated that the non-detection of a PBL is due to genuine lack of a BLR (e.g., Tran et al., 2011). Others have suggested that the non-detection is due to obscuration effects, rendering the detection of PBLs as difficult or impossible, even with deep near-IR (NIR) spectro-polarimetry (Alexander,

41 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

2001). Using a statistically complete IRAS 60 µm selected type-2 AGN catalog, Heisler et al. (1997) investigated the relationship between the detectability of PBLs and the torus inclination angle. They showed that only AGN with a low torus inclination angle have high detection rate of PBLs compared to those with high inclinations. This result strongly suggests that PBLs could be obscured when there is an edge-on view through the torus and/or nuclear obscuration in the host galaxies. In addition to the optical spectro-polarimetry, X-ray observations suggest that there is a weak evidence showing different absorption in two types of type-2 AGN. Gu et al. (2001) found that the AGN with PBL have slightly lower column density (NH) than those without PBL. Similarly, Lumsden et al. (2004) showed that the detection rate of PBL decreases as a function of NH, suggesting the absorption effect by dusty torus could play a role of the detectability of PBL in AGN. To understand the role of the obscuration by the torus in type-2 AGN and the detectability of PBLs, knowing the torus geometry/morphology and properties is crucial. In recent years much progress has been made toward understanding the torus geometrical structure. Thanks to the improvement of computing power, more physically realistic “clumpy” torus models have been coded by several authors (Hönig & Kishimoto, 2010; Hönig et al., 2006; Nenkova et al., 2002, 2008a,b; Schartmann et al., 2008; Stalevski et al., 2012). These models readily reproduce high spatial resolution nuclear NIR to mid-IR (MIR) spectral energy distributions (SEDs) and spectra of AGN with a compact torus of < 10 pc radius (e.g., Alonso-Herrero et al., 2011; Hönig et al., 2011; Lira et al., 2013; Nikutta et al., 2009; Ramos Almeida et al., 2009, 2011a). On the other hand, traditional smooth torus models (Efstathiou & Rowan-Robinson, 1995; Pier & Krolik, 1992, 1993) had difficulties to describe the variety of nuclear SEDs of nearby AGN (e.g., Alonso-Herrero et al., 2003; Asmus et al., 2011; Gandhi et al., 2009; Ichikawa et al., 2012a). Still, no one knows the true morphology of the tori until the future observations resolve the torus image because SEDs alone sometimes fail to resolve the degeneracy of the two models (e.g., Feltre et al., 2012). In this Chapter, under the assumption that the tori follow the distribution explained by clumpy torus model, we discuss how the precise modeled torus morphology plays a key role in the probability of the detection of PBL by fitting the clumpy torus model to our series of high-spatial resolution IR SEDs. We use 21 high spatial resolution MIR spectra in combination with NIR to far-IR (FIR) imaging sample, which is one the largest high spatial resolution NIR imaging+MIR spectroscopy+FIR imaging sample set of AGN in the local universe that is currently available. These high spatial resolution SEDs afforded by 8 m class telescopes minimize contamination of the MIR torus spectra from surrounded diffuse MIR emission from warm dust and/or stellar emission, and hence we are able to construct the highest fidelity torus

42 3.2 Observations

Table 3.1: Properties of the Sample

(lit) Name z d Slit/Size Type Group NH log Lbol b/a AV i Ref (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) NGC 1365 0.0055 18 0.35/31 Sy1.8 Type-1 23.6 42.9 0.5 <5 ··· (A1,B1,B1,··· ) NGC 4151 0.0033 13 0.36/23 Sy1.5 Type-1 22.8 43.7 0.71 ······ (A9,A9,··· ,··· ) IC 4329A 0.016 65 0.75/240 Sy1.2 Type-1 21.8 43.6 0.28 ······ (A10,A9,··· ,··· ) NGC 7469 0.016 66 0.75/240 Sy1 Type-1 20.7 45.1 0.72 ······ (A9,A9,··· ,··· ) NGC 1068 0.0038 15 0.36/26 Sy2 HBLR > 25 45.0 0.85 ··· 60–90 (A2,A9,··· ,A9) NGC 2110 0.0078 31 0.36/54 Sy2 HBLR 22.5 43.9 0.74 5 40 (A9,A9,A9,A9) MCG -5-23-16 0.0085 34 0.75/120 Sy2 HBLR 22.2 44.4 0.46 >6 53 (A9,A9,A9,A9) NGC 3081 0.008 32 0.65/100 Sy2 HBLR 23.9 43.8 0.8 ······ (A3,B2,··· ,··· ) NGC 3227 0.0039 17 0.75/62 Sy2 HBLR 22.2 43.4 0.68 ······ (A11,A9,··· ,··· ) Circinus 0.0014 4 0.60/12 Sy2 HBLR 24.6 43.6 0.44 9 60–90 (A8,A9,A9,A9) NGC 5506 0.0062 25 0.36/44 Sy2 HBLR 22.4 44.2 0.30 ≥11 40 (A9,A9,A9,A9) IC 5063 0.011 46 0.67/150 Sy2 HBLR 23.3 44.5 0.68 7 ··· (A2,A9,A9,··· ) NGC 7582 0.0053 21 0.75/76 Sy2 HBLR 22.7 43.3 0.42 8,13 ··· (A9,A9,A9,··· ) NGC 7674 0.029 118 0.75/430 Sy2 HBLR >25 45.0 0.91 ∼3–5 ··· (A9,A9,A9,··· ) NGC 1386 0.0029 11 0.31/17 Sy2 NHBLR >25.0 42.9 0.4 ··· 65,85 (A2,B2,··· ,C1) NGC 3281 0.011 43 0.35/73 Sy2 NHBLR 24.3 44.6 0.4 ······ (A4,B1,··· ,··· ) Cen A 0.0018 3 0.65/11 Sy2 NHBLR 23.7 44.0 0.4 14.0 ··· (A5,B2,A9,··· ) NGC 5135 0.014 59 0.70/200 Sy2 NHBLR >25.0 44.4 0.7 ······ (A2,B2,··· ,··· ) NGC 5643 0.004 16 0.35/29 Sy2 NHBLR 23.8 42.7 0.9 ······ (A6,B5,··· ,··· ) NGC 5728 0.0094 40 0.35/69 Sy2 NHBLR 23.6 44.5 0.6 ······ (A7,B6,··· ,··· ) NGC 7172 0.0087 35 0.36/61 Sy2 NHBLR 22.9 43.8 0.46 ······ (A2,A9,··· ,··· ) Sample properties. The sample is divided into three subgroups with type-1/HBLR/NHBLR respectively from top to bottom. (1) object name; (2) redshift; (3) luminosity distance (Mpc) gathered from literature for the case of nearby sources. Within the sample of González-Mart´ınet al. (2013), for NGC 1365, NGC 1386, NGC 1808, NGC3081, NGC 3281, and Cen A, the values of distance to the galaxies have been taken from Ramos Almeida et al. (2009). For NGC 5643, the distance has been taken from Guainazzi et al. (2004). For the sample of Alonso-Herrero et al. (2011), we gathered them from Alonso-Herrero et al. (2011). For the other sources, we calculated the distances by −1 −1 using cosmological parameter H0 = 75km s Mpc ; (4) slit width (arcsec) / physical size (pc); (5) Seyfert class of AGN. (6) Sub group of AGN. Type-1 represents type-1 AGN (Sy 1 to Sy 1.9) based on optical spectroscopy. HBLR represents type-2 (Sy2) AGN with hidden broad line region signs, and NHBLR represents type-2 AGN without any published hidden broad line regions signs. (7) hydrogen column density; (8) logarithm of bolometric luminosity (erg/s) which is taken from Alonso-Herrero et al. (2011); González-Mart´ın et al. (2013). We use a typical bolometric correction of 20 (Elvis et al., 1994). (9) The axial ratio; the ratio of the minor to major axis of the host galaxies. All information is taken from González-Mart´ın et al. (2013); (10) Foreground extinction in the unit of mag; (11) inclination angle of the torus. Levenson et al. (2006) derives the viewing angle of accretion disk of NGC 1386 and we here assume that the accretion disk and the torus are located in the same plane; (12) References of column (6), (7), (10), and (11). “··· ” denotes no reference. References. (A1) Alonso-Herrero et al. (2012a); (A2) Tran (2001); (A3) Moran et al. (2000); (A4) Nicastro et al. (2003); (A5) Alexander et al. (1999); (A6) Gu et al. (2001); (A7) Tran (2003); (A8) Wang & Zhang (2007); (A9) Alonso-Herrero et al. (2011); (A10) Véron-Cetty & Véron (2006); (A11) Imanishi (2002); (B1) Tueller et al. (2008); (B2) Marinucci et al. (2012); (B3) Brightman & Nandra (2011); (B4) Itoh et al. (2008); (B5) Guainazzi et al. (2004); (B6) Goulding et al. (2012); (C1) Levenson et al. (2006)

SED. Therefore we can probe the detectability of the PBL with the least amount of host galaxy contaminations than ever before.

3.2 Observations

3.2.1 The Sample

Our principal motivation in this study is to investigate whether the torus model morphology plays a major role in the chance of PBL detections. To achieve this, precise constraints of torus parameters is crucial. Therefore, high spatial resolution (∼ 0.3–0.7 arcsec) NIR and MIR imaging observations of AGN are crucial to disentangle the emission components from the host galaxy components and obtain torus SEDs with the least contamination as possible. In addition

43 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

to this, for achieving the optimal constraints of the torus parameters, the data set of MIR high spatial spectroscopy, which will trace the peak of torus SEDs and the 9.7 µm silicate feature, is crucial (see Ramos Almeida et al., 2014). We compiled the nearby AGN sources from the MIR samples of González-Mart´ın et al. (2013) (21 sources) and Alonso-Herrero et al. (2011) (13 sources) as both samples already compiled the currently available data set of high spatial N band spectroscopy. Of the 34 sources, 5 sources are in common in both samples, therefore the total number of the two catalogs is 29 sources. We further set the criteria for survey inclusion that the objects must also have at least one similarly high spatial resolution NIR (1–5 µm) measurement, as the NIR bands significantly help to constrain the torus parameters (Ramos Almeida et al., 2014) and those with high spatial resolution are crucial to disentangle the torus emission and the host galaxy to obtain the optimal torus NIR SED. 22 out of 29 sources fulfilled this criterion. We also removed NGC 1808 from this study as controversy remains to whether it hosts AGN or ultra-luminous X-ray sources in the galactic center due to the low X-ray −1 luminosity log L2−10 keV = 40.4 erg s (Jiménez-Bailónet al., 2005; Scarrott et al., 1993). We summarize the properties of the 21 sources in Table 3.1. (lit) Our sample spans AGN bolometric luminosities taken from the literature (Lbol ) in the range (lit) −1 −1 of log Lbol =42.7–45.1 erg s (see Table 3.1), with a mean value of 44.0 erg s . This value is fairly consistent with that of magnitude-limited Seyfert catalogs (Ho et al., 1997; Maiolino & Rieke, 1995). This suggests that our sample could be representative of AGN and their tori at least in the local universe, although the sample is not complete.

3.2.2 New Observations

We obtained N (Si2 ﬁlter; the central wavelength with λc = 8.73 µm and 50% cutoﬀ range of

∆λ = 0.39 µm) and Q (Qa ﬁlter; λc = 18.06 µm and ∆λ = 0.76 µm) band high spatial imaging data of NGC 5135 and NGC 5643, newly observed by T-ReCS (Program ID GS-2012A-Q-43, PI: Nancy Levenson). The standard MIR chop-nod technique was employed for the observations. The data reduction was conducted by using Redcan (Gonz´alez-Mart´ınet al., 2013).

3.2.3 Published Data from the Literature

To minimize the effect from the contaminating components in the host galaxies, we used the available high spatial resolution NIR and MIR imaging data sets from the literature. We collected the estimated values of the nuclear NIR to MIR emission when available. The compiled data originate from both ground-based and space telescope such as VLT/NACO and HST /NICMOS high spatial resolution observations. The only exception in our samples is NGC 5728, whose NIR nuclear flux is available by only estimated K band nuclear flux from the

44 3.2 Observations

2MASS (Peng et al., 2006). We set this data only as upper limit. All the NIR flux information is tabulated in Table 3.2 (from the column 2 to 6). We used the Spitzer/IRS 30 µm continuum fluxes of Deo et al. (2009) for those galaxies in our sample as upper limits in our fits because the star formation component already dom- inates at around 20 µm, and beyond 30 µm the component completely overwhelms the AGN torus emission in most cases of AGN (Netzer et al., 2007). This case, FIR data with large aperture are dominated with star formation component. Therefore, we did not include FIR points to the sources. However, we additionally included Spitzer/IRS spectroscopy data only when the AGN spectral turnover at 20–30 µm can be clearly detected in Spitzer/IRS Spec- troscopy. This suggests that the emission from AGN torus emission is dominated even in the large aperture size of Spitzer/IRS (Alonso-Herrero et al., 2012b; Hönig et al., 2014). Only two sources fulfilled the criterion (IC 4329A and MCG -5-23-16). For those galaxies, we collected available Herschel/PACS data. All the FIR flux information is tabulated from column 9 to 11 in Table 3.2. We apply the uniform error estimates for the analysis in this study. The errors were estimated using the prescription given by Alonso-Herrero et al. (2012a). Here we give a brief outline of these assumptions. For the NACO AO observations, we used 20% in J band and 15% in the HKLM band. For the other ground-based observation data, we applied 30% for J band, 25% for H and K band, and 20% for the L band. These errors include the photometric error, the background subtraction uncertainty, and the uncertainty from estimating the unresolved flux. M band fluxes were always used as upper limits due to the difficulties of estimating the unresolved component. For the NICMOS observations, we used 20% for the J band, 20% for the H and K band. For the N and Q band,we use 15% and 25% errors, respectively.

45 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN 97 (A7,A7) 45 (A7,A7) . . 0 0 /PACS photometry of et Meléndez < < m Ref 79 45 . . µ (A1) Carollo et al. (2002) (A2) Quillen et 1 1 Herschel < < 015 3 (A1,B1) . 8 (A7,A7) . m 70 58 (A2,B1) 64 (A7,A7) 98 (A7,A7) 03 (A2;A6, B2) 10 (A2,B1) 89 (A7,A7) 05 (A7,A7) 7683 (A7,A7) (A7,A7) . 0 ...... µ 0 12 1 3 0 3 1 1 1 3 4 ± < 30 References. < < < < < < < < < < 52 . 8 (A2,B2) 73 . . 8 . 278 (A4,B1) 5400 (A7,A7) 800 340 (A7,A7) 46 (A6,B1) 54 360 658 (A5,B1) 275 3200 (A7,A7) 550 204 114 140 220 132 (A7,A7) 57 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 2 92 . . 17 883 . 0 818 1 457 3 1110 5 231 . . . 454 218 . . 75 184 1500 21800 95 1450 44 561 198 3200 150 1 76 1350 30 15 78 48 1100 57 527 840 12800 . 8 107 2630 139 135 2200 30 22 53 3 12 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 31 36 . . List of Photometry 4 633 HBLR 4 165 . 340 10000 21 384 285 5600 NHBLR . ± ± ± 207 355 198 294 108 518 530 900 61 139 < < < < < < Table 3.2: 0 4 141 0 6 . . . . 0 0 320 6 30 710 57 1900 44 . . 140 2270 16 14 11 20 84 506 ± ± ± ± 325 1320 210 1014 33 47 ± ± ± ± ± < 5 0 0 < < . . . < < . The unit of flux is mJy from column (2) to (8), and Jy from column (9) to (11). All data at column 9 are taken 8 7 79 7 96 1 53 . 0 925 . . . 8 1 200 3 9 380 93 103 86 30 . . . . . 2 . 10 . 3 2 1 25 100 920 1 12 290 5 17 6 2 2 . 0 1 7 ± ± ± ± ± ± ± ± ± ± ± ± ± ± 7 3 0 8 . . . < 7 4 . . . BayesClumpy 0 445 4 178 0 102 6 18 . . 07 56 04 83 72 19 1 4 5 12 . . 0 68 1 23 0 80 8304 203 147 33 7 68 34 93 10 ...... 8 1 4 3 7 101 15 0 0 0 0 0 10 4 1 8 . . 0 0 0 0 0 0 1 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0 0 0 0 3 . . < < 3 5 3 7 2 . . . 66 22 77 ...... 0 98 25 5 2 40 0 53 33 3 26 4 ...... 2 14 104 0 3 3 72 0 60 4 0 0 . . m continuum fluxes of Deo et al. (2009) as upper limits. All data at column 10 and 11 are taken from JHKLMNQ 0 1 ± ± ± ± ± ± ± µ 8 1 3 . < < 25 . . . /IRS 30 Spitzer NGC 7172 NGC 3281Cen A 1 1 Name (1)NGC 1365NGC 4151 69 NGC 1068NGC (2) 2110 MCG -5-23-16 9 1 8 (3) (4)NGC 1386 (5) (6) 0 (7) (8) (9) (10) (11) (12) IC 4329A 50 NGC 7469 16 NGC 5728 NGC 5643 NGC 5135 NGC 3227 11 NGC 3081NGC 7674 0 1 Circinus NGC 7582 11 NGC 5506IC 5063 13 0 NIR to FIRfrom band list used as inputs for al. (2014). andal. are (2001); used (A3) as GallianoAlmeida upper & et limit. Alloin al. The (2008); (2009); references (A4) (B2) from Simpson Our NIR (1998); newly to (A5) observed MIR Meisenheimer data. band et fluxes al. are (2007); tabulated (A6) at Peng column et 12. al. (2006); (A7) Alonso-Herrero et al. (2011); (B1) Ramos

46 3.3 Application of Torus Model

3.2.4 Subsample

To examine the torus model morphological properties of different AGN populations, we divide the sample into subgroups based on whether or not the source has HBLR signs in previously published observations. We first divide the sample into type-1 and type-2 AGN. Although Seyfert 1.8/1.9 are very ambiguous objects (e.g., see Elitzur et al., 2014, for the details), here we define type-1 as AGN which have at least one broad emission lines in their optical spectra. Therefore, we consider Seyfert 1 to 1.9 as type-1 AGN and Seyfert 2 as type-2 AGN. Next, we divide the type-2 AGN into the AGN with or without any published polarized BLR detections in the optical band and/or HBLR detections in the NIR band (Non-HBLR; we call NHBLR hereafter). We use Marin (2014), who compiled almost all the previously published polarization information of nearby AGN. These spectro-polarimetric data are taken from several large surveys including the infrared-selected sample of Heisler et al. (1997), the far-infrared flux limited sample of Lumsden et al. (2001), the distance-limited sample of Moran et al. (2000, 2002), and the heterogeneous optical- and MIR selected sample of Tran (2001, 2003). Mainly the spectro-polarimetric observations were conducted by small or medium size telescopes (up to 4 m-class), while only NGC 3081 has been confirmed to have a HBLR features with the Keck 10 m telescope (Moran et al., 2000). Therefore, we should note that some HBLR AGN could contaminate the subgroup of NHBLR in the cases where the BLR is below the signal-to-noise afforded by the 4 m class telescope observations (see Ramos Almeida et al., in prep.). Some sources have currently no published spectro-polarimetric data, but have clear broad emission lines in NIR wavelengths. These sources are MCG-5-23-16, NGC 2110, and NGC 7582 (Alonso- Herrero et al., 2011). All the references used for dividing the sample into each subgroup are indicated in column 12 in Table 3.1. Finally, the sources in this study are categorized into three groups (type-1, HBLR, and NHBLR; see column 6 in Table 3.1). The sample contains 4 type-1, 10 HBLR, and 7 NHBLR AGN.

3.3 Application of Torus Model 3.3.1 Clumpy Torus Model

We ﬁt the clumpy torus models of Nenkova et al. (2008a), known as CLUMPY, to the data using a Bayesian approach (BayesClumpy; Asensio Ramos & Ramos Almeida 2009). Here we describe the six free CLUMPY model parameters used for the SED ﬁtting and the model set-up, which are listed in Table 3.3. The torus clumps are distributed in a smooth, rather than sharp, toroidal-shaped boundary of angular width σ. The inner radius (rin) of the torus is set

47 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

by the location of the dust at the sublimation temperature (Tsub ∼ 1500 K). This is computed

using the AGN bolometric luminosity Lbol(AGN)

L (AGN) 0.5 r = 0.4 bol pc. (3.1) in 1045 erg s−1

The torus has a radial extent (Y ) deﬁned by Y = rout/rin, where rout is the outer radius of the

torus. The average number of clouds along the line of sight (NLOS) at a viewing angle i is set as (90 − i)2 N = N exp − , (3.2) LOS 0 σ2

where N0 is the average number of clouds along the radial equatorial ray. NLOS allows us to

derive the escape probability of photons from the AGN central engines (Pesc). In the CLUMPY dust distribution, the classification of type-1 or type-2 AGN depends on whether or not there is a clump along the line of sight, which is a function of the viewing angle of the torus, the number of clumps and the torus width. This is different from smooth torus models, for which the classification of an AGN as type-1 or type-2 is solely determined by the viewing angle. The escape probability of photons passing through the torus at a given viewing angle (i) can be calculated as

−NLOS Pesc ∼ e . (3.3)

In the CLUMPY model, the radiative transfer equations are solved for each clump and thus the calculations depend on the clump distribution within the torus, the optical depth of each clump, and also its dust composition. Here we assume each clump has the same optical depth

(τV ), which is defined at the optical V band. The CLUMPY model applies a standard cold oxygen-rich interstellar medium dust, which is called OHMc dust (Ossenkopf et al., 1992). The torus clumps are distributed as a power law with index q as a function of radius, N(r) ∝ r−q. In addition to these six physical parameters, we add two additional parameters to be fitted or fixed. The first parameter is the foreground extinction (AV ), unrelated to the torus. Some authors demonstrated that some AGN have an extremely deep 9.7 µm silicate absorption feature which cannot be reproduced solely by the torus obscuration (Alonso-Herrero et al., 2003, 2011; González-Mart´ın et al., 2013; Goulding et al., 2012). They suggested that dust in inclined host galaxies can contribute significantly to the observed SED and silicate feature absorption. 10 out of 21 sources are inclined galaxies with low minor-to-major axis ratios (b/a ≤ 0.5; see Table 3.1). Therefore, some portion of the observed SED is accounted from by cool foreground dust extinction. Alonso-Herrero et al. (2011) discussed this issue and concluded that for AV ≥ 5, the effects of foreground extinction cannot be ignored for reproducing the silicate 9.7 µm

48 3.3 Application of Torus Model

Table 3.3: Free Parameters of the BayesClumpy

Parameters Parameter range (1) (2) Torus radial thickness (Y ) [5, 30] Torus angular width (σ) [15◦, 70◦] Number of clouds along an equatorial ray (N0) [1, 15] Index of the radial density proﬁle (q) [0, 3] Viewing angle (i) [0◦, 90◦] Optical depth of each cloud (τV ) [5, 150] Torus radial thickness Y is deﬁned as Y = rout/rin, where rout is the outer radius and rin is the inner radius. The −q cloud distribution between rout and rin is parameterized as r .

feature. We gathered available values of foreground AV from the literature and compiled them in column 10 in Table 1. The other additional parameter accounts for the multiplicative factor that has to be applied to match the ﬂuxes of a given model to an observed SED. Deriving this (mod) factor enables us to calculate the model AGN bolometric luminosity Lbol (Nenkova et al., (mod) (lit) 2008b). As shown by Alonso-Herrero et al. (2011), Lbol reproduces well the values of Lbol (mod) for Seyfert galaxies, and therefore, in the following we will refer to Lbol as the bolometric luminosities of the sample studied here.

3.3.2 Other Important Torus Parameters

By combining the derived output parameters of the CLUMPY model, we can derive other important torus morphological parameters as the torus outer radius rout, deﬁned as:

rout = rinY pc. (3.4)

We can also calculate the torus scale height H as:

H = rout sin σ pc. (3.5)

Finally, we define the “geometrical” torus covering factor, which is unaffected by the viewing angle, and it is defined by integrating the AGN escape probability over all angles (Nenkova et al., 2008a). This can be written as Z π/2 CT = 1 − Pesc(β) cos(β)dβ, (3.6) 0 where β = π/2 − i. Considering that our motivation is to characterize the intrinsic torus morphology, the “geometrical” torus covering factor is more relevant here than the apparent covering factor.

3.3.3 BayesClumpy and Modeling Details

The CLUMPY database currently contains more than 5 × 106 models. Therefore, when ﬁtting the models to the observations, inherent degeneracies have to be taken into account. We then

49 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

use the BayesClumpy fitting tool (Asensio Ramos & Ramos Almeida, 2009), as it does a fast synthesis of the CLUMPY SEDs. In the last version of BayesClumpy the inference over the model parameters can be done either using neural network interpolation or multilinear interpolation in the full database. After running several tests, Ramos Almeida et al. (2014) concluded that the latter interpolation produces more robust results. Therefore, here we use linear interpolation, which results in slight differences in the fitted parameters (within 1σ for the majority of the fits) for the 13 galaxies that were modeled by Alonso-Herrero et al. (2011) using the neural network interpolation and subsequently re-fitted in this paper. BayesClumpy can be used to fit photometry and/or spectra in a Bayesian scheme, carrying out inference over the model parameters for observed SEDs. This way we can specify a-priori information about the model parameters. Here we consider uniform prior distribution in the range of each parameter, as summarized in Table 3. The prior distribution of inclination angle (i) is fixed from previous observations if available in the literature, following the same approach as in Alonso-Herrero et al. (2011). From the objects in our sample taken from González-Mart´ın et al. (2013), NGC 1386 has two possible inclination angles 65◦ and 85◦ (Levenson et al., 2006). Thus, we use a uniform prior in the range 60◦–90◦ for this source. From the galaxies taken from Alonso-Herrero et al. (2011), we use the same inclination angle constraints they employed, which are compiled in column 11 in Table 3.1. We finally include the AGN direct emission (i.e. a broken power-law) which is defined in Eq. (13) of Nenkova et al. (2008a) to the SED for type-1 AGN, in order to reproduce the flatter slope of the NIR band.

3.4 Results and Discussions

3.4.1 Infrared SEDs with BayesClumpy Fitting

The result of the fitting process of the IR SEDs are the posterior distributions for the six parameters that describe the model (defined in Table 3.3), the foreground extinction and the multiplicative factor needed to match the SED fluxes. However, we can also translate the results into corresponding spectra. Figures 3.1, 3.2, and 3.3 show the observed SEDs and nuclear MIR spectra (black filled dots) of the galaxies with the best fit results overlaid (blue solid lines), based on the inference done with BayesClumpy. The fitted models correspond to those described by the median of the posterior distribution of each parameter. All the derived torus parameters obtained from BayesClumpy are presented in Table 3.4. Some SEDs show smaller Q band fluxes compared to the model spectra. This effect is

50 3.4 Results and Discussions

Figure 3.1: Model fits for type-1 AGN. - The filled dots are the SED photometric data and the black line is the MIR spectrum. The upper limit points are shown as arrows. The solid blue lines are the models computed with the median value of the probability distribution of each parameter. The blue shaded areas indicate the range of models compatible with a 68% confidence interval. For the details on the calculation of the median values, see Chapter 3.3.

51 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

Figure 3.2: Model fits for HBLR AGN. - The filled dots are the SED photometric data and the black line is the MIR spectrum. The upper limit points are shown as arrows. The solid blue lines are the models computed with the median value of the probability distribution of each parameter. The blue shaded areas indicate the range of models compatible with a 68% confidence interval. For the details on the calculation of the median values, see Chapter 3.3.

Figure 3.3: Model fits for NHBLR AGN. - The filled dots are the SED photometric data and the black line is the MIR spectrum. The upper limit points are shown as arrows. The solid blue lines are the models computed with the median value of the probability distribution of each parameter. The blue shaded areas indicate the range of models compatible with a 68% confidence interval. For the details on the calculation of the median values, see Chapter 3.3.

52 3.4 Results and Discussions

Table 3.4: Fitted torus model parameters from SED + spectroscopy data

(mod) Galaxy σtorus YN0 q τV i CT log Lbol rin rout H [deg] [deg] [erg/s] pc pc pc +5 +3 +2 +0.3 +42 +19 +0.08 +0.8 +0.005 +0.2 +0.1 NGC 1365 19−2 24−4 9−3 0.3−0.2 79−35 19−11 0.16−0.04 43.2−0.1 0.053−0.005 1.3−0.3 0.4−0.1 +1 +4 +1 +0.3 +8 +2 +0.03 +0.9 +0.007 +0.5 +0.2 NGC 4151 16−1 19−4 13−1 1.6−0.3 89−8 71−2 0.13−0.02 43.9−0.1 0.108−0.008 2.1−0.5 0.6−0.1 +1 +1 +1 +0.1 +1 +4 +0.04 +1.8 +0.002 +0.1 +0.0 IC 4329A 40−1 8−1 12−1 0.5−0.1 148−2 4−3 0.58−0.04 44.4−0.1 0.192−0.001 1.8−0.1 1.1−0.0 +2 +4 +1 +0.3 +12 +3 +0.05 +0.8 +0.020 +1.1 +0.4 NGC 7469 21−2 22−4 13−1 1.3−0.3 124−14 59−4 0.20−0.04 44.6−0.1 0.239−0.021 5.3−1.2 1.9−0.4 +8 +2 +3 +1.2 +3 +11 +0.06 +0.9 +0.015 +0.6 +0.3 NGC 1068 56−18 6−1 5−1 0.6−0.4 38−7 67−5 0.78−0.19 44.4−0.1 0.198−0.006 1.3−0.1 1.1−0.1 +8 +7 +2 +0.2 +2 +5 +0.05 +1.0 +0.003 +0.5 +0.4 NGC 2110 55−8 17−6 9−2 2.7−0.3 146−4 40−6 0.88−0.11 43.3−0.1 0.058−0.002 1.0−0.4 0.8−0.4 +5 +2 +2 +0.2 +3 +7 +0.03 +1.2 +0.004 +0.3 +0.3 MCG -5-23-16 58−8 20−1 7−1 2.1−0.1 144−6 48−3 0.85−0.05 43.9−0.1 0.114−0.003 2.3−0.2 2.0−0.3 +4 +9 +1 +0.2 +12 +13 +0.02 +1.0 +0.003 +0.5 +0.4 NGC 3081 62−5 11−4 12−1 2.6−0.4 98−11 66−19 0.96−0.03 43.1−0.1 0.043−0.003 0.4−0.1 0.3−0.1 +1 +1 +1 +0.1 +1 +5 +0.01 +1.0 +0.003 +0.1 +0.1 NGC 3227 57−1 20−1 13−1 0.0−0.0 147−2 6−4 0.94−0.02 43.0−0.1 0.041−0.003 0.8−0.1 0.7−0.1 +3 +2 +1 +0.3 +2 +8 +0.02 +1.0 +0.004 +0.2 +0.2 Circinus 64−5 16−1 12−2 0.4−0.2 35−2 37−7 0.96−0.03 43.4−0.1 0.062−0.003 1.0−0.2 0.9−0.1 +6 +3 +2 +0.3 +4 +4 +0.04 +1.0 +0.008 +0.5 +0.4 NGC 5506 48−3 16−2 10−3 0.2−0.2 79−5 32−1 0.79−0.05 44.0−0.1 0.130−0.007 2.1−0.5 1.6−0.3 +4 +7 +1 +0.2 +7 +7 +0.02 +1.0 +0.009 +1.6 +1.4 IC 5063 61−6 14−7 12−1 2.5−1.1 101−9 77−12 0.96−0.04 44.3−0.1 0.182−0.008 2.6−1.4 2.3−1.3 +3 +2 +1 +0.1 +7 +5 +0.02 +0.9 +0.005 +0.3 +0.2 NGC 7582 53−2 20−1 12−2 0.1−0.0 79−9 6−4 0.90−0.02 43.5−0.1 0.070−0.004 1.4−0.2 1.2−0.2 +13 +5 +3 +0.5 +9 +13 +0.17 +0.5 +0.068 +2.5 +1.4 NGC 7674 39−9 15−4 8−3 1.1−0.6 133−15 44−13 0.56−0.20 44.8−0.1 0.330−0.038 5.3−1.7 3.3−1.0 +7 +5 +2 +0.3 +4 +9 +0.05 +0.8 +0.002 +0.1 +0.1 NGC 1386 56−9 19−5 8−1 1.3−0.5 37−4 68−5 0.87−0.11 42.5−0.1 0.023−0.002 0.5−0.1 0.4−0.1 +1 +3 +1 +0.2 +3 +6 +0.01 +0.9 +0.010 +0.6 +0.5 NGC 3281 68−2 19−2 14−1 0.4−0.2 38−4 19−6 0.99−0.01 44.2−0.1 0.151−0.008 2.9−0.4 2.7−0.4 +10 +3 +2 +0.3 +11 +8 +0.10 +0.8 +0.002 +0.1 +0.1 Cen A 50−9 17−3 10−2 0.3−0.2 89−13 38−9 0.81−0.16 42.5−0.1 0.021−0.002 0.4−0.1 0.3−0.1 +3 +5 +1 +0.4 +5 +10 +0.01 +0.8 +0.007 +0.4 +0.4 NGC 5135 63−5 17−2 12−2 0.4−0.3 71−6 17−10 0.97−0.04 43.6−0.1 0.079−0.006 1.4−0.3 1.2−0.3 +4 +4 +1 +0.5 +11 +8 +0.02 +0.8 +0.004 +0.2 +0.2 NGC 5643 62−6 14−2 13−1 0.8−0.5 56−9 74−12 0.97−0.04 43.0−0.1 0.040−0.003 0.6−0.1 0.5−0.1 +2 +2 +1 +0.4 +7 +5 +0.01 +0.9 +0.004 +0.2 +0.2 NGC 5728 66−3 17−1 14−1 0.7−0.4 48−6 80−8 0.99−0.01 43.4−0.1 0.063−0.004 1.1−0.1 1.0−0.1 +1 +1 +1 +0.1 +1 +3 +0.01 +1.2 +0.002 +0.1 +0.1 NGC 7172 69−1 29−1 14−1 0.0−0.0 20−1 50−3 0.99−0.01 43.4−0.1 0.064−0.002 1.9−0.1 1.8−0.1 Notes.— Torus model parameters derived from the ﬁts with BayesClumpy. Median values of each posterior distribution are listed with their corresponding ±1σ values around the median.

prominent when silicate 9.7 µm feature shows deep absorption. Although this may suggest that the model spectra still have a diﬃculty to reproduce the 18 µm silicate feature, the diﬀerence is only within a factor of 3 even in the maximum case (e.g., NGC 3281).

The SEDs of NGC 5506 and NGC 7172 show the NIR excesses compared to the model spectra as shown in Figure 3.3. NIR interferometric observations (Kishimoto et al., 2009, 2011) and NIR reverberation mappings (Kishimoto et al., 2007; Koshida et al., 2009) of Seyfert AGN suggest that the AGN torus has much smaller sublimation radius than expected from Eq. (3.1). Kawaguchi & Mori (2011) also showed in their model that if they add the rim darkening effect of accretion disk, the torus inner radius naturally connects to the outer disk. Then, NIR emission excess naturally explain. This is also shown in Stalevski et al. (2012) when they apply the rim darkening effect to produce the model SEDs. However, in this study, we did not include the SED of a hot dust component or apply such torus geometry including rim darkening effect for the fitting as this hot dust remains rather unconstrained. The further discussion on the possible origins of NIR excesses can be found in Alonso-Herrero et al. (2011).

53 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

Table 3.5: Torus model parameters from the global posterior distributions

(mod) AGN Type σtorus YN0 q τV i CT log Lbol rin rout H [deg] [deg] [erg/s] pc pc pc +6 +2 +1 +0.6 +23 +14 +0.06 +0.5 +0.048 +0.4 +0.3 All 56−22 18−6 12−4 0.5−0.5 81−43 45−32 0.88−0.38 43.3−0.4 0.066−0.030 1.2−0.8 1.0−0.6 +3 +3 +0 +0.5 +20 +9 +0.06 +0.3 +0.048 +0.4 +0.0 Type-1 19−3 19−10 12−2 0.7−0.5 113−31 52−48 0.18−0.06 43.9−0.8 0.144−0.090 1.6−0.4 1.0−0.6 +4 +2 +1 +1.2 +34 +9 +0.04 +0.2 +0.054 +0.8 +0.6 HBLR 56−8 17−8 11−4 0.8−0.8 98−60 43−32 0.88−0.14 43.7−0.7 0.066−0.024 1.2−0.4 1.0−0.3 +2 +2 +0 +0.2 +17 +17 +0.02 +0.1 +0.006 +0.8 +0.3 NHBLR 64−11 18−3 13−3 0.4−0.4 43−11 48−28 0.96−0.10 43.2−0.8 0.060−0.042 0.8−0.4 1.0−0.6 Notes.— Torus parameters from the global posterior distributions of each subgroup.

3.4.2 General Torus Properties of Total Sample

The motivation of this chapter is to obtain a quantitative description of the torus morphology using the CLUMPY torus models. To take full advantage of the data employed here, we apply a hierarchical Bayesian approach to derive information about the global distribution of the CLUMPY parameters for a given subgroup. We use a generalized beta distribution as the prior for each parameter (given that they are defined in closed intervals) and learn the hyperparameters of the prior using importance sampling (e.g. Brewer & Elliott 2014). This allows us to derive the posterior distribution for each parameter taking into account all the observed data that belongs to an AGN subgroup. In Table 3.5 we report the median parameters of the posterior distributions for the total sample, as well as for each subgroup. Note that the posterior distribution of each parameter can be derived only in the circumstance that the prior distribution is the same for all the sources. This is not the case for the inclination angle because we use the constrained prior distributions for some sources as described in Chapter 3.3. For the inclination angle, we derive the median parameters of the individual galaxy fits within the each subgroup. +0.30 Our results show that the covering factor has the value of CT = 0.75−0.41, which is larger than 0.5. In a unified model scheme, this suggests that the number of objects that could be observed as type-2 AGN will be larger than that of type-1 AGN. This is in excellent agreement with the observational type-2 AGN ratio in our sample (17 type-2 AGN out of 21 sources, ∼ 0.8), and the average ratio from the literature at ∼ 70% (Schmitt, 2001) in the local universe. This suggests that our sample could be a representative of local AGN. Based on interferometry observations, Kishimoto et al. (2011) reported a typical torus half- light radius of ∼ 1 pc for local AGN at MIR wavelengths. Our derived torus outer radius +0.6 rout = 1.4−0.8 pc is consistent with the interferometry results, although a bit larger. This is understood because we consider that here we are considering colder dust within the torus than that traced by the MIR interferometry, as we include 20 µm data in our fits. These colder clumps will be generally located at larger radii, which explain the value of rout obtained from

54 3.4 Results and Discussions

Table 3.6: Results of KLD test for each parameter among each subgroup

AGN Type σtorus YN0 q τV Type-1 vs HBLR 3.13 1.26 0.10 0.42 0.10 Type-1 vs NHBLR 2.22 0.22 0.05 0.20 4.83 HBLR vs NHBLR 1.78 1.63 0.24 0.53 1.56 Notes.— KLD is calculated for the global posterior distribution of each parameter among the subgroups. Values larger than 1 are shown in bold.

the global posterior distribution. We also note that our value of rout does not include dust emitting in the far-infrared (FIR). The torus outer radius including cooler dust would be larger still than the value of rout ∼ 1.4 pc. Further studies including Herschel, SOFIA, and/or ALMA observations will help constraining the extent of the FIR-emitting dust. Indeed, some pilot studies already showed the importance of FIR data to trace cooler dust (Garc´ıa-Burillo et al., 2014; Ramos Almeida et al., 2011b).

1.4 6 2.0 1.2 5 1.0 1.5 4 0.8 3 0.6 1.0 2 0.4 0.5 1 Distribution 0.2 Distribution Distribution 0.0 0 20 30 40 50 60 70 0.0 5 10 15 20 25 30 2 4 6 8 10 12 14 σ[ ◦ ] Y N0

1.2 2.0 3.0 1.0 1.5 2.5 0.8 2.0 0.6 1.0 1.5 0.4 1.0 0.5 0.2

Distribution 0.5 Distribution Distribution 0.0 0.0 20 30 40 50 60 70 0.0 2 4 6 8 10 12 14 5 10 15 20 25 30 σ[ ◦ ] Y N0

0.7

40 0.6 0.5 30 0.4 20 0.3 0.2 10 Distribution Distribution 0.1 0 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 20 40 60 80 100 120 140 q τV

25 0.35 0.30 20 0.25 15 0.20 10 0.15 0.10 5 Distribution Distribution 0.05 0 0.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0 20 40 60 80 100 120 140 q τV

Figure 3.4: Posterior Distribution of each parameter. - Histograms of each physical parameter discussed in Chapter 3.4.3. Top panel of each ﬁgure represents the histogram of the whole sample. Blue/red/green ﬁlled color represents the histogram of Type-1/HBLR/NHBLR, respectively.

3.4.3 Distribution of Torus Parameters

The key science direction of this chapter is to investigate the torus morphology quantitatively by deriving torus parameters for each subgroup and comparing them. The median values of the model parameters ﬁtted to our nuclear IR SEDs are reported in Table 3.4. The median values of the global posterior distributions of the free parameters σ, Y , N0, q, and τV for each subgroup are compiled in Table 3.5. As discussed in Chapter 3.4.2, for the inclination angle i

55 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

14 10 12 150 8 10 6 8 100

4 6 4 50 2 Distribution Distribution Distribution 2 0 0 0 0 2 4 6 8 10 0 1 2 3 4 5 6 7 8 0.0 0.2 0.4 0.6 0.8 1.0 rout [pc] H [pc] CT

6 7 140 5 6 120 5 4 100 4 80 3 3 60 2 2 40 1 Distribution Distribution Distribution 1 20 0 0 0 0 2 4 6 8 10 0 1 2 3 4 5 6 7 8 0.0 0.2 0.4 0.6 0.8 1.0 rout [pc] H [pc] CT

Figure 3.5: Posterior Distribution of derived parameter. - Same as in Figure 3.4, but for the rout, H, and CT parameters.

we derive the median values from the individual galaxy fits for each subgroup. Figure 3.4 shows the global posterior distributions of each physical parameter for the different subgroups. Black, blue, red, and green histograms show the parameter distribution of all, type-1, HBLR, and NHBLR sources, respectively. As shown in Figure 3.4, for some of the parameters, the global distributions are similar for the three subgroups, but others are clearly different. In order to quantify these differences, we follow the same approach as in Ramos Almeida et al. (2011a). They used the Kullback-Leibler divergence (KLD; Kullback & Leibler, 1951) to show that the joint posterior distributions of type-1 and type-2 AGN were quantitatively different. The KLD takes into account the full shape of two posterior distributions to compare them. When the two distributions are identical, the value is KLD = 0 and, the larger the KLD value, the more different the two distributions. Ramos Almeida et al. (2011a) concluded that if KLD > 1.0, the two posteriors can be considered significantly different. We calculate the KLD values for the global distributions of each torus parameter among the three groups. The values are reported in Table 3.6.

We find significant differences for the parameters σ, Y , and N0 between type-1 AGN and HBLR AGN. The differences in these parameters between type-1 and type-2 AGN were already reported in Ramos Almeida et al. (2011a) with larger significance, but based on fits to NIR and MIR imagings only, where spectroscopic data were not included. Besides, they did not consider information from spectropolarimetry data, as we are doing here. Therefore, we confirm the results of Ramos Almeida et al. (2011a) after including N band spectroscopy to the IR photometry, which is crucial to constrain the six torus parameters (Alonso-Herrero et al., 2011;

Ramos Almeida et al., 2014). We also find that the parameters σ, Y , and N0 of HBLR and NHBLR AGN are significantly different. Considering the average values in Table 3.5, the tori

of NHBLR AGN have larger σ, larger Y , and larger N0 than those of HBLR AGN. There are various possible interpretations for the diﬀerence in σ among the subgroups. One possibility is that the smaller σ could be due to larger AGN luminosities (receding torus

56 3.4 Results and Discussions

model; Lawrence, 1991; Ricci et al., 2013). However, as shown in Table 3.5, the differences of median bolometric AGN luminosities among the subgroups are within the uncertainties. Therefore, we consider that the effect of the AGN luminosity is negligible in this study. Another possible interpretation is a selection bias in the optical type-1/type-2 AGN selection. Ramos Almeida et al. (2011a) and Elitzur (2012) discussed that AGN classification would depend on the distribution of the obscuring material; type-1 AGN would be preferentially selected from lower-obscuration AGN, while type-2 AGN (HBLR and NHBLR) from higher-obscuration AGN. This could be partly producing the differences in σ that we found for type-1 and type-2 AGN, but correcting this effect quantitatively is extremely difficult and beyond the scope of this study.

3.4.4 Distribution of Covering Factor

As shown in Chapter 3.3.2., we can derive physical parameters of the torus model by combining the model parameters. The individual values of these physical parameters are reported in Table 3.4, and those obtained from the global posterior distribution for each subgroup are shown in Figure 3.5 and Table 3.5. An interesting comparison can be done between the geometrical covering factor of the torus

(CT; described in equation 3.6) of the different subgroups and the average column densities derived from X-ray data (NH). NHBLRs have the largest column densities, with an average value −2 −2 of log NH ∼ 24.0 cm (i.e. Compton-thick), followed by HBLRs, with log NH ∼ 23.4 cm , and −2 type-1s, with log NH ∼ 21.8 cm . Based on hard X-ray (50–200 keV) observations of nearby AGN obtained with INTEGRAL, Ricci et al. (2011) reported differences between the column 23 −2 densities of type-1 and type-2 AGN. Type-1 and “lightly obscured” AGN with NH ≤ 10 cm have the same X-ray reflection component, with reflection amplitude R ∼ 0.4. On the other 23 −2 hand, “mildly obscured” AGN (NH ≥ 10 cm ) show a clearly stronger X-ray reflection component with R ∼ 2.2, suggesting that the central engine of “mildly obscured” AGN would be covered by an X-ray reflection wall. Our results are in good agreement with Ricci et al. (2011) if we consider the CT and NH values for each subgroup. The type-1 AGN in our sample fall in the “lightly obscured” AGN in their study, and indeed they show small covering factors

(CT ∼ 0.15), suggesting small torus X-ray reflection solid angle. The HBLR and NHBLR AGN subgroups would fall in the “mildly obscured” AGN category, and we found large covering factors for them (CT ∼ 0.90 and 0.97 respectively), suggesting a larger X-ray reflection component (see also Ricci et al., 2014). Figure 3.6 shows a schematic illustration of the torus geometrical differences among type-1 (left), HBLR (middle), and NHBLR (right).

57 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

Type-1 HBLR NHBLR

1 Figure 3.6: Schematic illustration of the torus geometry - Schematic illustration of the torus geometry for type-1 AGN (left), HBLR (middle), and NHBLR (right). The diﬀerence in color intensity between the two bottom panels shows the diﬀerence in optical depth of the clumps τV, where darker color is larger τV. The orange region represents the BLR. The green area represents the media where the incoming BLR emission is scattered and then polarized. The blue solid arrows represent the path of BLR photons and the blue dashed arrows represent the path of the polarized BLR photons. The observer is assumed to be on the left side of the torus with an inclination angle of 58◦, 43◦, and 49◦, respectively (see Table 3.5). The only photons scattered along these lines of sight are shown.

3.4.5 Torus Morphological Diﬀerences between HBLR and NHBLR

Here we focus on the differences between the modeled tori of HBLR and NHBLR. In the case of HBLR, we obtain smaller σ values than for NHBLR, which is equivalent to larger torus opening angles (90◦ − σ). This means that HBLR objects have a larger scattering region. (see middle panel of Figure 3.6). The scattering region (shown schematically as a filled green bar) can be larger due to the larger opening angle of the torus and it would produce a relatively higher amount of polarized flux from the BLR, which can then be observable from our line of sight. In the case of NHBLR, we obtain larger σ values than for HBLR. This means that the probability that scattered radiation from the BLR can be blocked is higher than for HBLR −2 objects. This is also in agreement with the larger value of log NH ∼ 24.0 cm estimated from X-ray observations of the NHBLR objects in our sample. To summarize, as shown in the bottom panel of Figure 3.6, the chance to observe scattered (polarized) flux from the BLR is reduced by the double effect of (a) less scattering of the flux from the BLR (due to the reduced scattering area) and (b) more obscuration between the observer and the scattering region. Therefore, the classification of an AGN as HBLR or NHBLR is probabilistic, and it would depend on the intrinsic properties of the torus, in particular of σ. This could be a reasonable explanation for the lack of hidden (polarized) BLR in some type-2 objects1. However, we note that the classification of the galaxies as HBLR and NHBLR is mainly

1 A similar explanation for the lack of HBLR detection in ∼40% of type-2 objects, based on the distribution of dust within the torus and its inclination being not as simple as predicted by the uniﬁed model, was shown in the talk by C. Ramos Almeida at the Polarization & active galactic nuclei workshop held 2012 October 16-17 at the Royal Observatory of Belgium.

58 3.4 Results and Discussions

based on spectropolarimetric observations from 3–4 m telescopes, and some of the NHBLR could be then misclassiﬁed. Therefore, further higher sensitivity spectropolarimetry observations of NHBLR AGN with 8 m class telescopes such as Subaru/FOCAS and/or VLT/FORS2 are highly encouraged to search for the HBLR in those AGN (Ramos Almeida et al. in prep.).

3.4.6 Inclination Angle Eﬀect on Detectability of HBLR in Type-2 AGN

8 S A R

I 6

h t i w

5 NHBLR

2 4 f

/ HBLR 0 6 f

0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 f60/f25 with BayesClumpy

Figure 3.7: Color-color plot obtained from IRAS and BayesClumpy - Plot comparing the f60/f25 colors measured from IRAS photometry and extrapolated from the clumpy torus model ﬁts. Blue circle/red square/green diamond represents type-1, HBLR, and NHBLR AGN, respectively. The larger symbol represents the average value of each subgroup. The horizontal line at f60/f25 = 4 is the boundary between HBLR and NHBLR AGN according to Heisler et al. (1997), where HBLR AGN have f60/f25 < 4 and NHBLR AGN f60/f25 > 4.

The dependence of HBLR detection in type-2 AGN on torus inclination angle is still a matter of debate. This idea arises from the observational trend that the IRAS 60 µm to 25 µm color ratio (f60/f25) of HBLR AGN is f60/f25 < 4 on average while that of NHBLR AGN is f60/f25 > 4 (Heisler et al., 1997; Lumsden et al., 2001). Several authors have suggested that this trend is due to the inclination angle of the torus: the cooler and outer dust within the torus blocks the warm and inner hot dust for edge-on views, producing the high f60/f25, while the warm inner dust can be seen from the face-on views, reducing the value of f60/f25. Following this idea, type-2 AGN with low f60/f25 would tend to be detected as HBLR AGN due to the more face-on view of the torus and vice versa. Here we can take advantage of the torus models ﬁtted to our SEDs, which are available for each AGN and shown in Figures 3.1, 3.2, and 3.3. The wavelength range covered by the models allows us to calculate f60/f25 color ratios for each source. We have also compiled f60/f25 color

59 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

ratios from IRAS for comparison, which probe much larger scales than our torus SEDs. Out of

21 sources, we obtained 17 IRAS f60/f25 colors with good quality of ﬂuxes (Fqual = 3, which is the highest quality)1.

Figure 3.7 shows the relationship between the f60/f25 ﬂux ratios obtained from IRAS and

those from the torus model SEDs. The averages IRAS f60/f25 ﬂux ratio for type-1, HBLR, and NHBLR AGN are 4.1 ± 2.5, 3.7 ± 2.0 and 6.4 ± 2.4, respectively, showing the previously mentioned correlation for HBLR and NHBLR AGN, although with large error bars. However, when we compare the values obtained from the torus model SEDs, which exclude contamination from the host galaxy, we ﬁnd that they are very similar for the three groups and smaller than

the IRAS colors (f60/f25 = 0.63 ± 0.07, 0.80 ± 0.48, and 0.75 ± 0.18 for type-1, HBLR, and NHBLR AGN respectively).

These results show that the differences of IRAS f60/f25 among the three subgroups are not produced from the torus dust. A similar result was reported by Alexander (2001), but using X- ray observations. Although the standard deviations of the average values of the IRAS colors are large for the three subgroups, one possible explanation for the cooler IRAS colors of NHBLR in comparison with those of HBLR AGN could be dust emission from stronger starbursts in their host galaxies. This is in good agreement with previous results showing that highly obscured AGN tend to have higher star formation activity in their host galaxies (Castro et al., 2014; Goulding et al., 2012; Ichikawa et al., 2012a,b, 2014a). Therefore, larger obscuration from the torus in NHBLR AGN (as shown in Figure 3.6) and higher star formation activity in the host galaxy could be somehow coupled. For example, using three-dimensional hydrodynamical simulations, Wada & Norman (2002) showed that starbursts and supernovae within the central 100 pc of host galaxies help lifting up the torus, suggesting that high star formation activity could influence the scale height of the torus. Comparing the nuclear and overall star formation activity of AGN with the torus obscuration is crucial to find out if they are coupled(e.g., Esquej et al., 2014; Ichikawa et al., 2014a; Imanishi et al., 2011a). Knowing the star formation activity of such AGN host galaxies is urged to unveil how star formation and torus obscuration are coupled (e.g., Ichikawa et al., 2014a; Imanishi et al., 2011a).

3.5 Conclusions

We constructed 21 infrared torus-dominated SEDs with high spatial resolution N band spectroscopy and Q band imaging obtained from Gonz´alez-Mart´ınet al. (2013) and Alonso-Herrero et al. (2011), and with NIR high spatial resolution imaging from the literature. By performing

1 See Beichman et al. (1988) for the deﬁnition of Fqual in the IRAS catalogs. False detections may be included when Fqual < 3.

60 3.5 Conclusions

SED ﬁtting using clumpy torus models and a Bayesian approach we derived torus parameters such as the torus covering factor (CT), the torus inner and outer radius (rin and rout), and the torus scale height (H). We divided the sample into subgroups based on whether or not they are optically type-1, type-2 with observational hidden broad line regions sign (HBLR), and type-2 without any observational broad line region signs (NHBLR). Our results are summarized as follows:

1. We obtained a quantitative description of the torus geometry and intrinsic properties. We

found that the median torus outer radius for the whole sample rout = 1.4 pc is consistent with the results from MIR interferometry observations.

2. We found that the tori of Type-1 AGN have smaller σ, Y , NH, and CT than those of HBLR and NHBLR. Moreover, the tori of NHBLR are thicker and therefore have higher

CT than those of HBLR. These diﬀerences the torus properties of HBLR and NHBLR AGN would make it more diﬃcult to detect hidden BLR in NHBLR.

3. Combining f60/f25 colors obtained from IRAS photometry and from torus model SEDs,

we showed that the low f60/f25 measured fof HBLR using IRAS data are not due to a more face-on inclination of torus, but rather to star formation activity in their host galaxies.

61 3. CLUMPY TORUS MODELING OF POLARIZED AND NON-POLARIZED BLR AGN

62 4

AKARI IRC 2.5–5 µm Spectroscopy of Infrared Galaxies over a Wide Luminosity Range

4.1 Introduction

Dust emission gives us crucial information to understand both the history of cosmic star formation in galaxies and that of supermassive black hole (SMBH) growth in galactic centers. Intense star formation produces a great amount of dust, which makes its activity invisible in the ultraviolet band but visible in the infrared band. Similarly, the central engines of active galactic nuclei (AGN) are surrounded by dusty “tori” (Krolik & Begelman, 1986). Since optical and ultraviolet lights are very easily obscured by the torus, the most complete search for AGN can be made by detecting hard X-rays from the central engine or infrared emission from the heated dust. Utilizing X-ray data, as discussed in Chapter 1.7, the fraction of Compton-thick AGN increases with redshift from z = 0 to z > 1. In addition, Ueda et al. (2014) suggested that AGN co-moving luminosity density has a peak around z ∼ 2. These kinds of facts imply that the level of dust-obscuration of AGN may be correlated to the star formation activity, which has a peak at z ∼ 2 (Hopkins & Beacom, 2006). Although, the hard X-rays are useful to 24.5 −2 search for obscured AGN, extremely heavily buried AGN with NH > 10 cm are diﬃcult to be detected due to ﬂux attenuation by repeated Comptonization even at energies above 10 keV (e.g., Brightman & Nandra, 2011; Ikeda et al., 2009). Candidates of galaxies that host such buried AGN are infrared galaxies as discussed in Chapter 1.7. Their bolometric luminosities are dominated by the infrared emission, suggesting that very luminous heating sources are surrounded by dust and then the heated dust re-emits in the infrared band. The hidden energy sources are believed to be either SB, AGN, or both.

63 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

Disentangling the energy sources of infrared galaxies is crucial to unveil the history of dust obscured star formation and SMBH growth in the universe (Goto et al., 2010; Le Floc’h et al., 2005; Magnelli et al., 2011; Murphy et al., 2011). The former studies using Spitzer were compiled in Chapter 1.7. Infrared 2.5–5 µm spectroscopy is another powerful tool to study optically-elusive buried AGN in U/LIRGs. One advantage of this band is that dust extinction is much lower than in the optical band (Nishiyama et al., 2009) and is similar to that in the 5–13 µm band (Lutz et al., 1996). Another great advantage is that SB and AGN activity can be distinguished based on the spectral features. First, strong polycyclic aromatic hydrocarbon (PAH) emission, a pure star formation tracer, is located at 3.3 µm in SB galaxies, while pure AGN exhibit PAH-feature free spectra due to the eﬀects that 1) the X-ray emission destroying the PAHs, and 2) the strong continuum emission originating from AGN hot-dust diminishes this feature (Imanishi & Dudley, 2000; Moorwood, 1986). SB galaxies generally show large 3.3 µm PAH equivalent widths of

EW3.3PAH ∼ 100 nm, which never go down below 40 nm. Thus, objects with EW3.3PAH ≤ 40 nm may be classiﬁed as galaxies that harbor buried AGN. In SB/AGN composite galaxies, strong PAH emission can be observed because the AGN cannot destroy the PAH molecules located in the outer (r > 10 pc) region from the central engine due to shielding by dust and gas. However, strong AGN continuum makes EW3.3PAH smaller, well below 40 nm. Using this method, many authors reported the strong sign of buried AGN in ULIRGs (Imanishi et al., 2006; Risaliti et al., 2010; Sani et al., 2008) and LIRGs (Imanishi et al., 2008, 2010). Another key feature in Γ the 2.5–5 µm band is the continuum slope Γ (Fν ∝ λ ). SB galaxies have blue continuum slopes (Γ ∼ 0) in this band due to the contribution of the stellar photospheric continuum. Conversely, galaxies with AGN have hot-dust emission heated by AGN, which produces a red continuum (Γ ≥ 1). This method was ﬁrst reported by Risaliti et al. (2006) and successfully applied to ULIRGs (Risaliti et al., 2010; Sani et al., 2008) and LIRGs (Imanishi et al., 2008,

2010). Imanishi et al. (2010) combined the two methods (EW3.3PAH ≤ 40 nm and/or Γ > 1) to ﬁnd buried AGN from a sample of U/LIRGs showing no AGN signatures in the optical spectra. They found that the fraction of buried AGN increases with infrared luminosity in the range of 11 13 10 L ≤ LIR ≤ 10 L . In this chapter, we expand the studies of U/LIRGs by Imanishi et al. (2010) into a lower 10 infrared luminosity range, by including normal infrared galaxies (IRGs) with 10 L ≤ LIR < 11 10 L , where comprehensive studies are still missing due to the limited sample. This is because L/IRGs have spatially more extended infrared emission (> 1 kpc, corresponding to >several arcseconds at z ∼ 0.05) compared with ULIRGs (Imanishi et al., 2011b; Soifer et al., 2001), and therefore slit spectroscopy with ground based telescopes could miss their extended

64 4.2 Targets

emission. Slit-less spectroscopy of Infrared Camera (IRC; Onaka et al., 2007) on board AKARI (Murakami et al., 2007) with 10 × 10 aperture opened a new window to probe such extended −1 −1 emission of L/IRGs. Throughout this chapter, we adopt H0 = 75 km s Mpc ,ΩM = 0.3, and ΩΛ = 0.7 for consistency with the previous publications (Imanishi et al., 2008, 2010).

4.2 Targets

1014

1013 sun 12

/L 10 IR L

1011

1010 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 z

Figure 4.1: Redshift distribution of our AKARI sample - Redshift versus infrared luminosity plot for the 22 infrared galaxies in our sample of this chapter.

Our motivation is to search for buried AGN from optically non-Seyfert infrared galaxies and investigate their properties over a wider range of infrared luminosity than in previous studies. An ideal sample can be selected from unbiased catalogs of infrared galaxies with various infrared luminosities at redshifts of z < 0.45, where 3.3 µm PAH is detectable in the AKARI /IRC 2.5– 5.0 µm spectra. To this end, we first gather U/LIRGs with high far-infrared fluxes from IRAS catalogs. Our ULIRGs are mainly selected from the bright ULIRG catalog by Klaas et al. (2001), consisting of 41 sources 25 out of which are optically non-Seyfert galaxies. These ULIRGs have 12 1 IRAS 60 µm fluxes ≥ 3 Jy and far-infrared (40–120 µm) luminosities L40−120 µm > 10 L . In addition to these sources, we also gather 54 fainter, local (z < 0.45) optically non-Seyfert ULIRGs from the literature to increase the sample size. The LIRGs are mainly taken from the LIRG catalog of Carico et al. (1988), which is based on the IRAS Bright Galaxy catalog. The catalog consists of 61 sources with the criteria of IRAS 60 µm fluxes ≥ 5.4 Jy and far-infrared 11 (40–400 µm) luminosities L40−120 µm > 10 L . We select 41 optically non-Seyfert galaxies

1 −1 −1 This luminosity is based on the paper of Klaas et al. (2001), where H0 = 50 km s Mpc . Therefore, in −1 −1 11.5 our study with H0 = 75 km s Mpc , some galaxies reach LIR = 10 L .

65 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

out of this sample. To further increase the number of L/IRG targets, we also utilize the catalog by Spinoglio et al. (2002). They compiled 76 ISO detected sources in the IRAS 12 µm galaxy catalog, which gathered all galaxies at Galactic latitudes of |b| > 25◦ down to the IRAS 12 µm ﬂux limit of ≥ 0.22 Jy (Rush et al., 1993). Since AGN tori have peak emission at the MIR band, 12 µm galaxy catalog could tend to include more AGN than in far-infrared selected samples. Indeed, a majority of the Spinoglio et al. (2002) catalog (64 out of 76 sources) are optically Seyfert galaxies. Therefore, we gather the 12 optically non-Seyfert starburst galaxies from it. In summary, we select a suﬃciently large number of non-Seyfert infrared galaxies (132 sources) as our “parent” sample to be followed-up with AKARI. The redshifts of all the sources in the above catalogs are distributed within z < 0.45.

4.3 Observations and Data Reduction

Observation was conducted with the IRC infrared spectrograph (Onaka et al., 2007) on board AKARI (Murakami et al., 2007). The spectra in the 2.5–5 µm band were taken with the NG grism mode. This mode achieves an effective spectral resolution of R ∼ 120 at 3.6 µm, which is sufficient for detecting and tracing the profile of 3.3 µm PAH emission. The IRC has a 1 × 1 arcmin2 window with a pixel scale of 1.46 × 1.46 arcsec2. All the data were taken as a part of the AKARI mission program called “AGNUL” (PI: T. Nakagawa). The observation settings were same as those described in Chapter 3 of Imanishi et al. (2010). One to seven pointings were assigned for each source according to the brightness. The total on-source exposure time was ∼ 6 minutes per each pointing. We used IRCZ4 (phase 3; post liquid-He mission) observing mode, where one pointing is composed of eight or nine independent frames (Onaka et al., 2007). This mode successfully removes cosmic-ray contaminations even for sources observed by a single pointing. For sources observed with multiple pointings, we combined all the data to achieve the best signal-to-noise (S/N) ratio after excluding the data sets with particularly bad quality. We often had to discard data observed in later epochs of AKARI ’s phase 3 operation due to the increasing background signal. Spectral analysis was conducted in a standard manner by using the IDL data reduction package for AKARI IRC spectra1. The dark-subtraction, linearity correction, and flat-field correction were performed using this IDL package. Some infrared galaxies with low luminosities 12 (LIR < 10 L ) have spatially extended emission. Hence, we varied the aperture size for spectral extraction according to the actual signal profile of each source. The background was estimated from both sides of the spectral dispersion direction of the target and was subtracted

1The software package is available through the AKARI data reduction webpage http://www.ir.isas.jaxa. jp/AKARI/Observation/. For the details of the data analysis, see Ohyama et al. (2007).

66 4.4 AGN/SB Spectral Decomposition and Buried AGN Diagnostics

from the source. The IDL package for AKARI also performed wavelength and absolute flux calibration. The accuracy of the wavelength calibration is ∼ 1 pixel (∼ 0.01 µm), and that of the flux calibration is ∼ 10% around the central wavelength of the spectra and up to ∼ 20% at the edges close to 2.5 µm and 5.0 µm. Out of the 132 sources in our parent sample, we observed 37 objects in total from 2008 June until 2010 February with AKARI. After the epoch, AKARI fell into the stage where the data highly suffered from the significant background emission. As the result, we could not observe the remaining 95 targets. Also, due to AKARI ’s Sun-synchronous polar orbit, higher visibilities were achieved for objects at higher ecliptic latitudes. These AKARI ’s observational constraints make it difficult for us to construct a uniformly flux-limited sample over the entire sky with AKARI. However, the observable targets are essentially randomly selected from the parent sample with no selection biases regarding the physical nature of the galaxies. Hence, we regard that our sample has no obvious bias. Next, we also set the criterion that the averaged S/N ratio for each spectral channel must be higher than 5 in order to apply our spectral deconvolution analysis. Among the 37 targets observed with AKARI, five sources were discarded due to their low S/N ratio and the other 10 sources were too faint to be detected. 14 out of 15 non-detected sources are fainter ULIRGs gathered from the literature. These sources have 60 µm flux = 1.1 Jy, which is roughly 10 times fainter than that of the detected sources (f60µm = 11.8 Jy). Supposing that their infrared spectral energy distribution does not change drastically from that of the detected sources, we can regard that the most of the non-detected sources are intrinsically too faint to be observed with AKARI /IRC. Figure 4.1 shows the redshift distribution of the finally selected 22 infrared galaxies.

4.4 AGN/SB Spectral Decomposition and Buried AGN Diagnostics

For understanding the properties of buried AGN in infrared galaxies, it is crucial to analyze the infrared spectra by decomposing SB and AGN components. The 2.5–5.0 µm band covered by AKARI IRC has unique advantages for detecting AGN signatures. As introduced in Chap- ter 4.1, one promising method for ﬁnding buried AGN is the strength (equivalent width) of the 3.3 µm PAH emission as proposed by Imanishi & Dudley (2000). Pure SB galaxies show generally strong 3.3 µm PAH emission with EW3.3PAH ∼ 100 nm because plenty of SB-produced UV photons excite PAH molecules (Mouri et al., 1990). By contrast, if an AGN resides in the galaxy, the PAH feature becomes weaker, EW3.3PAH < 40 nm (Imanishi & Dudley, 2000).

67 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

Table 4.1: Basic Information of Our Sample

Name ObsID z f12 f25 f60 f100 log LIR Components (1) (2) (3) (4) (5) (6) (7) (8) (9) ESO 286-IG19 1122176-001 0.043 0.28 1.90 12.71 10.58 12.00 2comp IC 5135 1120236-001 0.016 0.63 2.14 16.67 26.27 11.33 2comp IRAS 03068-5346 1122105-001 0.074 0.25 0.20 3.43 3.92 11.98 2comp IRAS 03538-6432 1122106-001 0.30 0.25 0.25 0.96 1.54 13.01 2comp IRAS 10494+4424 1122096-001 0.092 0.25 0.25 3.36 5.60 12.22 2comp IRAS 17028+5817 1122101-001 0.11 0.25 0.08 2.49 4.05 12.22 2comp IRASF 07353+2903 1120200-001 0.33 0.08 0.10 0.22 0.64 12.63 2comp MCG +02-04-025 1122123-001 0.031 0.37 1.46 11.13 10.29 11.67 2comp MCG +08-23-097 1122130-001 0.029 0.25 0.53 5.08 8.10 11.34 2comp MCG +10-19-057 1122132-001 0.031 0.40 1.92 11.35 10.81 11.70 2comp Mrk 1490 1122134-001 0.026 0.25 0.86 6.21 8.38 11.30 3comp Mrk 551 1122171-001 0.050 0.26 0.80 4.63 6.13 11.80 2comp Mrk 848 1122136-001 0.040 0.32 1.52 9.38 10.26 11.86 2comp NGC 2339 1120230-001 0.0074 0.53 2.11 18.96 32.24 10.69 3comp NGC 2388 1120231-001 0.014 0.49 1.99 16.21 23.09 11.15 3comp NGC 4102 1120232-001 0.0028 1.45 6.83 47.37 68.50 10.25 2comp NGC 4194 1122091-001 0.0083 0.86 4.36 22.79 25.94 10.88 3comp NGC 4818 1120234-001 0.0036 0.85 3.88 20.26 26.55 10.10 3comp NGC 520 1120229-001 0.0076 0.78 2.83 31.52 48.40 10.91 3comp NGC 6285 1122138-001 0.019 0.30 0.44 7.43 23.58 11.21 3comp NGC 838 1122173-001 0.013 0.59 1.83 0.40 17.94 10.74 3comp ZW 453.062 1122140-001 0.025 0.29 0.57 7.59 11.04 11.34 2comp Table 4.1 summarize the basic information of our targets. (1) object name: (2) observation ID of AKARI : (3) redshift: (4)–(7) 12, 25, 60, and 100 µm IRAS flux density in the unit of Jansky (Jy): (8) total infrared (8–1000 µm) luminosity 39 2 in units of solar luminosity (L ), calculated from LIR = 2.1 × 10 × D(Mpc) × (13.48f12 + 5.16f25 + 2.58f60 + f100) (Sanders & Mirabel, 1996): (9) number of continuum components adopted to fit each spectrum: “2comp” and “3comp” correspond to the model spectra given by equation (6) and (7) in Chapter 4.4, respectively. In a case of “3comp”, the main flux contribution originates from stellar and Hii components with little contribution from dust component. This is why for “3comp” sources the dust component (dashed-dot line) is hardly seen in Figure 4.2.

Γ Imanishi et al. (2008, 2010) also used the red continuum slope (Γ > 1 for Fν ∝ λ ) for finding buried AGN. In this paper, we develop the previous works not only to find buried AGN but also to quantitatively estimate their energy contribution by performing detailed spectral fitting with a continuum model consisting of a stellar component and dust components from AGN and/or SB. Emission and absorption line features are also included in the model.

4.4.1 AGN-heated dust component

The AGN emission in the near- to mid-infrared originates from the radiation reprocessed by hot dust in the torus. Recent studies of dusty torus models reproduce the observed nuclear mid-infrared emission well by assuming that the torus media have a clumpy structure (e.g., Alonso-Herrero et al., 2011; Nenkova et al., 2008a,b; Ramos Almeida et al., 2009, 2011a). However, some authors mentioned that there are diﬃculties in ﬁtting their near-infrared spectral energy distribution (SED) (Lira et al., 2013; Stalevski et al., 2012; Videla et al., 2013). This is due to the complicated degeneracies including a possible existence of an extra hot-dust (∼ 1500 K) component originating from the vicinity of the AGN (Kishimoto et al., 2011), and/or contamination from the host galaxies, and/or the extinction of the torus emission by interstellar matter in the host galaxies (e.g., Alonso-Herrero et al., 2011). In this work, we take

68 4.4 AGN/SB Spectral Decomposition and Buried AGN Diagnostics

the simplest approach to approximate the torus dust emission by a single blackbody component, considering our limited wavelength range. Presumably, the dusty torus has a continuous dust temperature distribution and produces emission peaked around ∼10 µm. Hence, our single blackbody ﬁt should trace its peak temperature at ∼300 K. Accordingly, we set a conservative upper limit on its temperature as Tdust < 800 K (Oyabu et al., 2011).

4.4.2 Stellar and Starburst Component

The stellar photospheric emission in the host galaxy contributes to the total radiation in the 2.5–5.0 µm spectral range. This component produces a decreasing continuum flux at ≥ 1.8 µm (Sawicki, 2002). We first determine the temperature T (stellar) by fitting the 2MASS J, H, and

Ks photometric data with a blackbody model, and fix the temperature at its best-fit value when analyzing the AKARI 2.5–5.0 µm spectra. For Mrk 551 and Mrk 848, we cannot find any 2MASS fluxes. Therefore, we use the J, H, and K photometric data from the literature (Spinoglio et al., 1995) instead of the 2MASS data. SB produces a large amount of dust, which generally has lower temperatures than those of the AGN-heated dust, typically T (dust) < 100 K (Sanders & Mirabel, 1996). As already suggested by Sawicki (2002) and many other studies, such cold dust with T (dust) < 100 K does not significantly contribute to the 2.5–5.0 µm spectra, and therefore we ignore the contribution from the SB dust emission. However, some SB galaxies show excess emission from very hot dust with a temparature of ∼ 103 K, requring at least two blackbody components. This very hot dust component of ∼ 103 K would not originate from the dusty torus, from which blackbody radiation of ∼300 K is generally expected (see Chapter 4.4.1.). One possible origin is the emission mainly heated by massive stars in Hii regions (Hunt et al., 2002; Lu et al., 2003; Siebenmorgen, 1993). If such emission is required from the data, we model it with a single blackbody with a temperature of THii ∼ 1000 K in a range of 800 < THii < 1200 K as suggested by Hunt et al. (2002).

4.4.3 Emission and Absorption lines

The 2.5–5.0 µm spectra of infrared galaxies contain various emission/absorption line features. The previous studies of the AKARI “AGNUL” program (Imanishi et al., 2008, 2010) reported three PAH lines at 3.3, 3.4, and 3.5 µm, hydrogen Brakett series at 4.05 µm Brα and 2.63 µm Brβ, and Pfund series at 4.65 µm Pfβ and 3.74 µm Pfγ as emission lines. They also reported absorption features at 3.1 µm by H2O ice, at 4.29 µm by CO2, and at 4.67 µm by CO. Be- cause PAH emission is believed to originate from solid state molecules, the line proﬁle becomes Lorentzian (Yamada et al., 2013). For simplicity, however, we model all the lines including those of PAH by Gaussian proﬁles, following Imanishi et al. (2010). The central wavelength and line

69 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

Table 4.2: Fitting Properties and AGN Signs Obtained from the AKARI 2.5–5.0 µm Spec- troscopy (stellar) (Hii) (dust) Name EW3.3PAH T T T AGN AGN (1) (2) (3) (4) (5) (6) (7) ESO 286-IG19 68.9 ± 5.2 2786 ± 94 ··· 358 ± 4 N Y IC 5135 47.3 ± 1.2 2929 ± 37 ··· 749 ± 20 N Y IRAS 03068-5346 90.4 ± 5.7 3148 ± 411 ··· 700 ± 82 N Y IRAS 03538-6432 31.9 ± 4.9 1475 ± 80 ··· 304 ± 15 Y Y IRAS 10494+4424 81.2 ± 4.8 2083 ± 144 ··· 415 ± 22 N Y IRAS 17028+5817 54.2 ± 4.5 2206 ± 199 ··· 366 ± 17 N Y IRASF 07353+2903 5.2 ± 3.9 1544 ± 154 ··· 249 ± 56 Y Y MCG +02-04-025 121.3 ± 5.0 2545 ± 86 ··· 349 ± 11 N Y MCG +08-23-097 41.4 ± 1.6 2612 ± 74 ··· 649 ± 64 N Y MCG +10-19-057 76.1 ± 3.0 2455 ± 68 ··· 708 ± 71 N Y Mrk 1490 59.7 ± 1.6 2568 ± 70 1128 ± 94 < 100 N N Mrk 551 41.5 ± 2.8 2545 ± 141 ··· 499 ± 13 N Y Mrk 848 93.9 ± 4.2 3142 ± 227 ··· 715 ± 43 N Y NGC 2339 42.3 ± 1.4 2968 ± 41 1199 ± 275 < 100 N N NGC 2388 55.3 ± 1.1 2715 ± 37 1177 ± 95 < 100 N N NGC 4102 38.7 ± 1.7 2971 ± 44 ··· 739 ± 27 Y Y NGC 4194 78.2 ± 1.8 2986 ± 46 1192 ± 68 < 100 N N NGC 4818 42.0 ± 1.3 3240 ± 61 1200 ± 217 < 100 N N NGC 520 54.2 ± 1.1 2897 ± 55 1149 ± 75 < 100 N N NGC 6285 59.9 ± 2.1 2940 ± 83 1173 ± 194 < 100 N N NGC 838 110.9 ± 1.2 2863 ± 41 1200 ± 185 < 100 N N ZW 453.062 50.9 ± 3.2 2725 ± 55 ··· 649 ± 71 N Y Table 4.2 summarizes the observed properties of the 22 sources obtained from the AKARI spectroscopy. (1) object name: (2) equivalent width of 3.3 µm PAH emission: (3), (4), (5) temperature of black body originated from stellar emission, Hii emission, and dust emission. See Chapter 4 for the detail: (6) AGN sign based on the PAH diagnostic. “Y” represents the source has buried-AGN sign (EW3.3PAH < 40 nm), while “N” represents no AGN sign (EW3.3PAH > 40 nm): (7) AGN sign based on the hot-dust diagnostic. “Y” represents the source with T (dust) > 200 K, while “N” represents the source with T (dust) < 200 K.

width are ﬁxed within a spectral resolution at appropriate values reported in the literature.

4.4.4 Total model spectra

We thus construct a spectral model consisting of the multiple continuum components over which the emission/absorption lines are superposed. For the continuum we consider three (stellar) blackbody components, (1) a direct stellar component fBB , (2) a dust component from (dust) AGN with a temperature below 800 K fBB , and (3) a very hot (800 K–1200 K) dust com- (Hii) ponent from Hii regions fBB . Each blackbody component has two parameters, normalization (C(stellar),C(dust),C(Hii)) and temperature (T (stellar),T (dust),T (Hii)), which, except for T (stellar), are left as free parameters in the spectral ﬁt. Previous studies suggest that emission from warm dust may be partly absorbed by outer cold dust (e.g., Imanishi et al., 2010; Risaliti et al., 2010). To take this into account, we apply an extinction correction by incorporating τ(λ) into the blackbody components. We adopt the dust extinction index of β = 1.75 (Draine, 1989; Hunt et al., 2002) as λ −1.75 τ(λ) = τ . (4.1) 0 µm

The normalization value τ0 can be estimated from the optical thickness of 3.1 µm ice (τice), (dust) which is calculated by using the deconvolved continuum from the dust (fBB ). Imanishi &

70 4.4 AGN/SB Spectral Decomposition and Buried AGN Diagnostics

Maloney (2003) estimated the relationship between the dust extinction AV and τice with

τice = 0.06AV (1 + f), (4.2) where f is the fraction of dust that is covered with an ice mantle. The maximum value of f is derived to be 0.3 in the core of U/LIRGs in Imanishi & Maloney (2003). From the relation between the optical thickness and extinction with τV = AV /1.08 and Eq. (4.2) with f = 0.3,

τ τ = ice = 11.87τ (4.3) V 0.06 × 1.3 × 1.08 ice is derived. By combining Eq. (4.1) and (4.3), at the V band (0.55 µm),

−1.75 τ(λ = 0.55 µm) = τ0 × 0.55 = 11.87τice (4.4)

⇐⇒ τ0 = 4.3τice (4.5)

(dust) (dust) −τ(λ) Accordingly, we model ths observed component as fBBobs = fBB (λ) × e , where τ(λ) = −1.75 4.3 × τice × (λ/µm) . First, we adopt only two continuum components of direct stellar emission and AGN/SB dust emission to ﬁt the observed spectra, because the contribution of dust emission from Hii regions is not always required from the data. This model (2 component model) is written as:

(stellar) (dust) X (line) fmodel(λ) = fBB (λ) + fBBobs(λ) + fi (λ), (4.6) i

(line) where fi (λ) represents each line’s proﬁle as described in Chapter 4.4.3. In some cases, (dust) mainly for sources showing little contribution from fBB , the 2-component model given by (Hii) equation (4.6) does not ﬁt the data well. Then, we add the fBB component to the above model:

(stellar) (Hii) (dust) X (line) fmodel(λ) = fBB (λ) + fBB (λ) + fBBobs(λ) + fi (λ). (4.7) i

We adopt this model (3 component model) only when the fitting result is significantly improved from the 2-component model with a > 95% confidence level on the basis of an F-test. The fitting model we adopt for each source is summarized in Table 4.1. As shown in Figure 4.2, all sources that require the three components are galaxies showing blue contina in their spectra (e.g., NGC 4818). As many authors (e.g., Imanishi et al., 2010; Risaliti et al., 2010) suggested, these galaxies do not show buried AGN signs but are more likely normal SB galaxies. This is consistent with our assumption that the extra hot (T ∼ 103 K) component should originate from Hii regions, not from the AGN torus.

71 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

ESO286−IG19 IC5135 IRAS03068−5346 IRAS03538−6432 flux (mJy) flux (mJy) flux (mJy) flux (mJy) 0 10 20 0 20 40 60 80 0246 024 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 wavelength(µm) wavelength(µm) wavelength(µm) wavelength(µm)

IRAS10494+4424 IRAS17028+5817 IRASF07353+2903 MCG+02−04−025 flux (mJy) flux (mJy) flux (mJy) flux (mJy) 024 024 0 0.5 1 1.5 2 0 5 10 15 20 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 wavelength(µm) wavelength(µm) wavelength(µm) wavelength(µm)

MCG+08−23−097 MCG+10−19−057 Mrk1490 Mrk551 flux (mJy) flux (mJy) flux (mJy) flux (mJy) 0 5 10 15 20 0 10 20 0 5 10 15 0 5 10 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 wavelength(µm) wavelength(µm) wavelength(µm) wavelength(µm)

Mrk848 NGC2339 NGC2388 NGC4102 flux (mJy) flux (mJy) flux (mJy) flux (mJy) 0 5 10 0 20 40 60 0 20 40 60 80 0 50 100 150 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 wavelength(µm) wavelength(µm) wavelength(µm) wavelength(µm)

NGC4194 NGC4818 NGC520 NGC6285 flux (mJy) flux (mJy) flux (mJy) flux (mJy) 0 20 40 60 80 0 50 100 0 50 100 150 0 5 10 15 20 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 3 3.5 4 4.5 wavelength(µm) wavelength(µm) wavelength(µm) wavelength(µm)

NGC838 Zw453.062 flux (mJy) flux (mJy) 0 10 20 0 50 100 3 3.5 4 4.5 3 3.5 4 4.5 wavelength(µm) wavelength(µm)

Figure 4.2: AKARI 2.5–5.0 µm spectra - AKARI 2.5–5.0 µm spectra (red dots with error bars) of the 22 infrared galaxies overplotted with the best-ﬁt models (blue solid curve). The black dashed, dashed-three-dot, dashed-dot curves represent the stellar, Hii, and AGN/SB dust components, respectively. Black dotted curves represent emission lines.

72 4.5 Results and Discussions

4.5 Results and Discussions 4.5.1 AKARI spectra and AGN diagnostics

Figure 4.2 displays the observed AKARI spectra of all the 22 sources (red dots with flux error bars) overplotted with their best-fit models (blue line). Each line/continuum component is also plotted in the figure. As noticed, a majority of sources show 3.3 µm PAH emission features. Table 4.2 and 4.3 summarize the fitting results. For finding buried AGN, we apply two AGN diagnostics: (1) PAH emission line and (2) torus-dust continuum. The PAH diagnostic is based on EW3.3PAH as discussed in Chapter 4.4; we identify a buried AGN if it shows a small 3.3 µm PAH equivalent width with EW3.3PAH < 40 nm. The torus-dust diagnostic is based on whether or not there is a contribution from a hot dust component from the torus; we set the criterion of identifying buried AGN as T (dust) > 200 K.

73 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE : (6) 1 − s 2 − erg cm 13 − IR 008 004 005 020 002 010 055 292 015 010 034 005 020 005 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 /L ± ± ± ± ± ± ± ± ± ± ± ± ± ± ··· ··· ··· ··· ··· ··· ··· ··· , respectively: (10) logarithmic 1 (dust) BB 031 053 035 228 010 038 366 308 085 055 180 020 329 085 − ...... L spectroscopy. (1) object name: (2), 03 0 01 01 01 02 01 02 0 01 01 02 0 01 03 0 02 0 02 0 02 0 62 0 04 0 03 0 07 0 03 0 03 0 01 0 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3PAH . ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3 AKARI L m 49 18 40 23 12 67 30 77 52 90 43 86 33 39 80 74 82 02 58 95 77 65 ...... µ log 41 40 41 41 41 40 41 41 m PAH flux in units of 10 µ 14 41 04 40 06 41 04 41 11 41 13 41 07 41 30 41 09 41 09 42 09 42 13 41 03 41 03 41 ...... 0 0 0 0 0 0 0 1 0 0 0 0 0 0 2.5–5.0 (dust) BB ± ± ± ± ± ± ± ± ± ± ± ± ± ± ··· ··· ··· ··· ··· ··· ··· ··· L 42 56 99 74 26 51 82 71 74 54 85 87 10 84 ...... log 43 42 43 44 43 43 44 45 44 44 45 43 45 43 AKARI 03 09 04 11 08 12 05 05 ) ...... 0 0 0 0 0 0 0 0 Hii BB ( ± ± ± ± ± ± ± ± L ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· 41 64 36 25 64 02 65 64 , respectively: (5) 3.3 ...... log 1 − s 2 01 05 43 02 42 02 43 03 43 05 43 01 03 43 02 43 03 04 43 02 01 01 01 07 04 03 02 06 01 01 ...... − 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± (stellar) BB L 49 93 54 46 07 07 40 30 74 58 77 07 02 42 35 66 70 66 60 80 64 62 ...... erg cm , and dust-torus blackbody component in units of erg s log 43 43 43 44 44 44 43 44 12 Hii − 07 44 03 43 03 44 06 44 04 44 06 44 21 45 12 44 08 44 07 44 05 44 ...... 0 0 0 0 0 0 0 0 0 0 0 03 44 09 45 02 44 . . . ice 0 0 0 ± ± ± ± ± ± ± ± ± ± ± ··· ··· ··· ··· ··· ··· ··· ··· τ < < < 32 20 18 08 39 27 38 30 02 20 13 ...... 26 33 31 15 56 0 16 0 07 17 19 11 0 09 06 04 0 06 0 13 0 01 0 02 0 03 1 02 05 0 21 12 0 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3PAH . : (11) ratio of the AGN blackbody luminosity to the total infrared luminosity. 3 11 43 10 55 10 1 Observed Properties Obtained from the 47 86 68 88 43 04 39 11 41 20 02 26 56 15 75 66 58 . . . . . f ...... − 1 8 5 2 11 10 15 14 9 8 6 2 20 13 5 2 10 1 3 1 5 1 52 3 4 0 4 0 6 0 21 1 15 0 2 0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ··· ··· ··· ··· ··· ··· ··· ··· (dust) BB 7 21 30 10 19 22 19 29 f 16 137 235 106 340 338 Table 4.3: 35 19 6 22 13 13 25 4 ) ± ± ± ± ± ± ± ± Hii ( BB ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· f 5442 181 206 6 26 6 9 5 4 4 , and AGN blackbody component in units of 10 1328 26 137 26 193 26 120 33 101 7 33 2 2 2 2 1 12 7 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Hii (stellar) BB f m PAH emission in units of erg s O ice: (7), (8), (9) logarithmic luminosity of the stellar, µ 2 NGC 4818NGC 520NGC 6285NGC 838ZW 453.062 1191 1052 120 370 245 NGC 4194 257 NGC 2388NGC 4102 534 1629 NGC 2339 526 Mrk 551Mrk 848 113 117 Mrk 1490 89 MCG +10-19-057 54 MCG +08-23-097 153 MCG +02-04-025 114 IRASF 07353+2903 15 IRAS 17028+5817 21 IRAS 10494+4424 25 IRAS 03538-6432 16 IC 5135IRAS 03068-5346 54 856 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Name ESO 286-IG19 112 Table 4.3 summarizes the observed properties including continuum fluxes and luminosities of the 22 sources obtained from the (3), (4) flux of the stellar, optical thickness of H luminosity of the 3.3

74 4.5 Results and Discussions

In the left panel of Figure 4.3, we present the two AGN diagnostic diagram applied for our sample, where ULIRGs/LIRGs/IRGs are plotted as pink circles/red triangles/brown squares, respectively. There is a clear boundary of distribution of the dust temperature around T (dust) ∼ 200 K. This supports our criterion that AGN should have high dust temperatures (T (dust) > 200 K) and are well distinguishable from SBs. Interestingly, many infrared galaxies with large (dust) PAH equivalent widths (EW3.3PAH > 40 nm) show high dust temperatures (T > 200 K). One reason could be that because of the large aperture of AKARI /IRC and less luminous nucleus emission in these galaxies, the spectra contain relatively large contribution of 3.3 µm PAH emission in the host galaxies where PAH molecules are not destroyed by AGN X-ray photons. This effect works to increase the PAH equivalent width, and hence the PAH diagnostic could miss buried AGN. By contrast, the hot-dust diagnostic is not subject to the aperture effect because the continuum emission can be properly decomposed by the spectral fit. In fact, the torus-dust AGN diagnostic can recover all the buried AGN identified by the PAH diagnostic; in other words, there is no source with a low dust temperature (T (dust) < 200 K) and a small

PAH equivalent width (EW3.3PAH < 40 nm) in our sample. This supports the superiority of using the hot-dust diagnostic for ﬁnding AGN completely.

140 140 ULIRG LIRG IRG 120 120

100 100

[nm] 80 [nm] 80

60 60 3.3PAH 3.3PAH

40 40 EW EW

20 20

0 0 100 200 300 400 500 600 700 800 1042 1043 1044 1045 1046 (dust) T [K] log LBB(dust) [erg/s]

Figure 4.3: Equivalent width of 3.3 µm PAH vs. dust temperature - Plot of equiv- (dust) alent width of 3.3 µm PAH emission (EW3.3PAH) versus dust temperature T (left panel) and infrared AGN luminosity (right panel). Pink circles, red triangles, and brown squares represent ULIRGs, LIRGs, and IRGs, respectively. In the right panel, we plot only the sources with T (dust) > 200 K because those with T (dust) < 200 K has a great temperature uncertainty, and therefore the luminosity is given only the small upper limit with 1040 erg s−1 (see the discussion in Chapter 4.5.1).

(dust) The right panel of Figure 4.3 shows the plot of EW3.3PAH versus dust luminosity (LBB ) tabulated in Table 4.2 and Table 4.3. In the ﬁgure, we plot only the sources with T > 200 K. This is because the 2.5–5.0 µm spectral range is not sensitive to the dust emission with low temperature T < 100 K. In this case, only an upper limit of the dust luminosity can be derived. (dust) (dust) This is the reason why the dust luminosity with T < 200 K is very small, LBB

75 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

1040 erg s−1. The figure shows that all the sources with T (dust) > 200 K have significant dust (dust) 42 −1 emission with LBB > 10 erg s . Thus, the alternative criterion for detecting AGN can be (dust) 42 −1 (dust) LBB > 10 erg s instead of T > 200 K . In summary, we conclude that the hot-dust AGN diagnostic is a better method for finding buried AGN in the AKARI IRC band.

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2 Buried AGN fraction Buried AGN fraction

0 0 10 10.5 11 11.5 12 12.5 13 10 10.5 11 11.5 12 12.5 13 log LIR/Lsun log LIR/Lsun

Figure 4.4: Buried AGN fraction as a function of infrared luminosity. - The left panel shows the result based on the 3.3 µm PAH AGN diagnostic, and the right panel shows that based on the torus-dust AGN diagnostic.

4.5.2 Buried AGN contribution as a function of infrared luminosity

One motivation of our study is to reveal how frequently buried AGN exist in infrared galaxies. Figure 4 shows the fraction of buried AGN as a function of infrared luminosity. The left and right panels show the results based on the PAH AGN diagnostic and on the torus-dust AGN diagnostic, respectively. As discussed above, the latter diagnostic is more complete than the former for finding buried AGN from infrared galaxies. Therefore, we discuss the result of the right panel hereafter. In total, we detect 14 buried AGN out of the 22 infrared galaxies without any AGN signs in the optical band. We also find a clear trend that the fraction of buried AGN increases with infrared luminosity. While only 17% (1 out of 6 sources) of the IRGs contain buried AGN, the U/LIRGs contain them almost ubiquitously (8/11 = 72% for LIRGs and 5/5 = 100% for ULIRGs). Our results for U/LIRGs are consistent with those obtained by Imanishi et al. (2010) within the statistical uncertainties. Although most of the IRGs are selected from the 12 µm galaxy catalog, which is sensitive to torus-dust emission, an AGN is detected only from one source out of the five targets. The high AGN fraction in U/LIRGs may be related with their high merger rates. There is strong observational evidence that most of U/LIRGs have experienced merging or strong galaxy- galaxy interactions (Veilleux et al., 2002). The mergers can induce strong SB within the galaxies,

76 4.5 Results and Discussions

which also make galactic gas fall into the galactic center and trigger AGN activity obscured by dust (Di Matteo et al., 2005; Hopkins et al., 2006; Mihos & Hernquist, 1996; Sanders et al., 1988). Sanders & Ishida (2004) show that infrared galaxies with signs of recent merger drastically 11.2 increase at LIR ∼ 10 L toward a higher luminosity range. This result is in accordance with 11 our ﬁnding that the buried AGN fraction drastically increases from LIR ∼ 10 L .

10-2 ULIRG w/ buried AGN LIRG w/ buried AGN LIRG w/o buried AGN IRG w/ buried AGN IRG w/o buried AGN 10-3 IR / L 10-4 3.3PAH L 10-5

10-6 1010 1011 1012 1013 1014 LIR [erg/s]

Figure 4.5: Ratio of L3.3PAH/LIR as a function of infrared luminosity. - Open circles/triangles/squares represent ULIRGs/LIRGs/IRGs with AGN signs, respectively. Filled triangle/square represents LIRGs/IRGs without AGN signs, respectively. Note that all ULIRGs in our sample are diagnosed to have buried AGN.

We conﬁrm the dependence of the AGN fraction on luminosity by investigating the ratio between the 3.3 µm PAH luminosity (L3.3PAH) and total infrared luminosity LIR. Pure SB −3 galaxies generally have a constant ratio L3.3PAH/LIR ∼ 10 . If an AGN exists inside the galaxy, the ratio becomes smaller because X-rays from the AGN destroy PAH molecules and the infrared continuum luminosity is increased by the torus-dust emission. Figure 4.5 plots

L3.3PAH/LIR as a function of infrared luminosity for our sample. As noted, it decreases with 11 infrared luminosity, especially at LIR > 2 × 10 L . We also ﬁnd that the sources with small

L3.3PAH/LIR ratios match to those with buried AGN signs, supporting the above idea. Note that there is a possibility that this trend may be caused by the destruction of small PAH molecules by major mergers (Yamada et al., 2013) rather than by AGN. Nevertheless, since −4 all the sources with L3.3PAH/LIR <∼ 10 show buried AGN signs, the conservative criterion −4 L3.3PAH/LIR < 10 could be phenomenologically used as another diagnostic for ﬁnding buried AGN.

4.5.3 Luminosity contribution of buried AGN to the total infrared luminosity

Our second motivation in this chapter is to determine the energetic importance of AGN in these infrared galaxies. As discussed in Chapter 5.5.1, the luminosity contribution from dust

77 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

0.3

0.25

0.2

0.15

0.1

0.05 Buried AGN contribution

0 10 10.5 11 11.5 12 12.5 13 log LIR/Lsun

Figure 4.6: Energy contribution of buried AGN to the total infrared luminosity as a function of infrared luminosity. -

(dust) blackbody component (LBB ) predominantly originates from AGN-heated dust, and the lumi- (dust) nosity contribution of SB-heated dust to LBB is negligible. Thus, using the best-fit blackbody parameters obtained by the spectral fit, we can estimate the total luminosity from AGN-heated dust and hence its ratio to the total infrared luminosity. Figure 4.6 shows the averaged fraction of AGN power to the total infrared luminosity for IRGs, LIRGs, and ULIRGs. As noticed, the AGN contribution increases with infrared luminosity, while the values are quite small, ∼ 0.9 ± 0.8% in IRGs, ∼ 7.4 ± 3.3% in LIRGs, and ∼ 19.1 ± 5.0% in ULIRGs. This suggests that the bulk of the infrared emission originates from SB, not from AGN in these galaxies. One caveat for this result is that we cannot completely exclude the possibility that the estimated (Hii) luminosity of LBB might be partially originated from the torus-dust if it has an extreme tem- (Hii) perature distribution. We thus calculate the averaged LBB /LIR ratio to check if it could have (Hii) a significant contribution. We find LBB /LIR ∼ 7%, suggesting that the main results above would not be affected, even if the observed very hot (> 1000 K) dust component were totally due to the AGN emission.

Another caveat for interpreting this result is the selection bias for our sample, which consists of optically non-Seyfert galaxies. Previous studies based on samples including optically Seyfert AGN found larger average AGN contribution to the total infrared luminosity, 10–15% in LIRGs (Alonso-Herrero et al., 2012a; Petric et al., 2011) and 15–40% in ULIRGs (Nardini et al., 2010; Risaliti et al., 2010; Veilleux et al., 2009).

Our results are in good agreement with those by Lee et al. (2012) within the errors, who obtained the AGN to total luminosity ratio of 6–8% in LIRGs and 11–19% in ULIRGs, using a similar sample to ours that preferentially includes optically non-Seyfert infrared galaxies. They calculated the AGN contribution by performing SED ﬁt with the Decompir package (Mullaney

78 4.5 Results and Discussions

et al., 2011). This method requires infrared SED data covering the far-infrared band, while our method only uses the AKARI 2.5–5 µm spectra. This supports that our simple assumption in the spectral model (ie, single blackbody for dust emission) works well in estimating the AGN luminosity. This simple method will have a great advantage in the era of JWST, because we can apply our method to high-z galaxies by observing the rest 2.5–5.0 µm band (5.0–10.0 µm at z = 1 and 7.5–15.0 µm at z = 2) with JWST /MIRI (5–28 µm). Other methods based on the Spitzer bandpass (5–35 µm) would have a diﬃculty in studying high-z objects; for instance, the observed spectral range corresponds to 15–105 µm at z = 2, which only the SPICA mission can cover. Thus, our method is achievable for studying infrared galaxies in the distant universe even before the launch of SPICA (from 2025–).

1010

109

) 8

3 10

107 /Mpc sun 106 (L IR 105 Ω

Total IR 104 ULIRG LIRG ULIRGAGN LIRGAGN 103 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 z

Figure 4.7: Comoving infrared luminosity density as a function of redshift. - Black filled diamonds represents the infrared luminosity density (ΩIR) of all galaxies, pink filled circles (ULIRG) (LIRG) that of ULIRGs (ΩIR ), and red filled triangles that of LIRGs (ΩIR ), taken from Goto et al. (2010, 2011). Pale-pink filled circles and brown filled triangles represent the AGN infrared luminosity density of ULIRGs and that of LIRGs estimated by our work.

Combining our result with the total infrared luminosity density produced by U/LIRGs in (LIRG) 6 3 (ULIRG) 5 3 the local universe (ΩIR ∼ 5.9×10 L /Mpc and ΩIR ∼ 1.4×10 L /Mpc ) (Goto et (AGN) al., 2011), we estimate the local AGN luminosity density in the infrared band (ΩIR ). We find (AGN) 5 3 (AGN) 4 3 ΩIR = 4.4 × 10 L /Mpc for LIRGs and ΩIR = 2.6 × 10 L /Mpc for ULIRGs. These values are about an order of magnitude lower than the estimate by Goto et al. (2011). This is because Goto et al. (2011) simply assumed that the infrared luminosity of galaxies classified as AGN entirely originates from AGN, while we quantitatively take into account the fraction of AGN contribution to the total infrared luminosity of infrared galaxies. Figure 4.7 shows the infrared luminosity density as a function of redshift derived by Goto et al. (2010, 2011) together with our results of AGN the infrared luminosity density. In the era of JWST, we can fill in the AGN infrared luminosity density at high redshifts thanks to their improved sensitivity in the

79 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

mid-infrared band.

4.5.4 Evolutionary track of IR galaxies

Comparison between the starburst and AGN activity gives crucial information to understand the process of galaxy-SMBH co-evolution. For optically selected Seyfert galaxies and QSOs, (AGN) many authors found that the luminosity of star formation (LSF) and that of AGN (Lbol ) are well correlated with each other (Imanishi et al., 2011a; Netzer, 2009; Oi et al., 2010). We (dust) (dust) can derive a SF luminosity as LSF = LIR − LBB . The AGN-heated dust luminosity (LBB ) (AGN) summarized in Table 4.3 can be converted to an AGN bolometric luminosity (Lbol ) by using two relations given in Gandhi et al. (2009) and Marinucci et al. (2012), respectively,

(AGN) log L2−10 keV = 0.90 log LMIR + 4.09 (4.8) (AGN) log Lbol ∼ log Ldisk = 1.18 log L2−10 keV − 6.68. (4.9)

Here Ldisk represents the intrinsic luminosity of the accretion disk integrated in the optical to (AGN) (dust) X-ray band. Assuming LMIR ∼ LBB , we obtain

(AGN) (dust) log Lbol ∼ 1.06 log LBB − 1.85. (4.10)

(AGN) Noted that the estimated bolometric AGN luminosity (Lbol ) derived in Equation (4.10) can be overestimated. The relations from Equation (4.8) to (4.10) are derived from the sample of non-hidden Seyfert AGN, while our sources are infrared galaxies with highly obscured (=hidden) AGN. Therefore, our sources in this study could have higher covering fraction of dusty-torus than those of Seyfert AGN. This means the ratio of mid-infrared over bolometric luminosity in our sources could be larger than those of Seyfert AGN. (AGN) Figure 4.8 shows the LSF versus Lbol relations obtained from our sample. The average luminosities of U/LIRG are also plotted in the same figure. The dotted line corresponds to the linear correlation obtained from optically selected Seyfert galaxies obtained by Netzer (2009). Interestingly, all the sources except two (IRAS F07353+2903 and ESO 286-IG19) are located above the Seyfert line. This result may reflect the difference in the evolutional stage of galaxies, from a pure SB phase (U/LIRGs) to an unburied AGN phase (Seyferts). As discussed in Chapter 4.5.2, U/LIRGs have recently experienced or are facing mergers, which result in strong

SB activity from the epoch T1 to T2 in Figure 4.8. The SB also triggers AGN activity, which continues from T2 to T3. Finally, the SB activity is gradually weakened due to the shortage of gas and feedback from the AGN, and the supply of matter from the host galaxy onto the central engine also decreases, from T3 to T4. In this framework of galaxy evolution, the evolutional stage of ULIRGs should be a later phase of mergers than that of LIRGs. The picture is in good

80 4.6 Conclusions

47 ULIRG LIRG IRG 46

45 [erg/s] SF 44 log L

42 42 43 44 45 46 47 log Lbol(AGN) [erg/s]

Figure 4.8: Evolutional Track of AGN. - (Left): Plot of SF Luminosity (LSF) versus (AGN) AGN bolometric luminosity (Lbol ). The dotted line represents the correlation line for optically selected Seyfert galaxies taken from Netzer (2009). All symbols and colors are the same as Fig- ure 4.3. (Right) A cartoon illustrating a scenario of AGN evolution from a pure SB galaxy to an unburied AGN phase. T1 represents the evolutional stage when SB is triggered by merger, T2 when AGN activity is induced by strong SB, T3 when SB activity starts to decrease due to the shortage of gas and feedback from the AGN, and T4 when both SB and AGN activities become weaker. agreement with observations that the morphology of ULIRGs are more compact than that of LIRGs, suggesting that ULIRGs have just finished the merger process while that of LIRGs is still on-going. This scenario suggests that infrared galaxies containing buried AGN could be an earlier evolutional stage of AGN, which will be evolved to normal (unburied) Seyferts/QSOs. (AGN) This scenario becomes much more reasonable if our bolometric AGN luminosity (Lbol ) is overestimated as discussed in the previous paragraph because all the points are shifted to the left in the figure. While our sample size is still small, the new AGN diagnostics developed in this paper can be applied to other sources. Further studies using a well-defined complete sample of infrared galaxies should be urged to complete our view of AGN evolution.

4.6 Conclusions

We obtained the AKARI 2.5–5.0 µm spectra of 22 infrared galaxies at z < 0.35 selected mainly from three infrared catalogs: bright ULIRG catalog (Klaas et al., 2001), IRAS 12 µm galaxy catalog (Spinoglio et al., 2002), and IRAS bright galaxy catalog (Carico et al., 1988; Soifer et al., 1987). This band includes unique spectral features for ﬁnding “buried” AGN, the 3.3 µm PAH emission line and red continuum spectra originated from AGN torus emission. We perform detailed spectral ﬁtting to these data by properly modeling the continuum and emission/absorption features. This enables us to decompose the continuum spectra into SB and AGN torus-dust components. The results are summarized as follows.

81 4. AKARI IRC 2.5–5 µM SPECTROSCOPY OF INFRARED GALAXIES OVER A WIDE LUMINOSITY RANGE

1. Using the AGN diagnostics based on the torus-dust continuum (T (dust) > 200 K or (dust) 42 −1 LBB > 10 erg s ) and 3.3 µm PAH emission strength (EW3.3PAH < 40 nm), we ﬁnd that 14 out of the 22 infrared galaxies have buried AGN. We conﬁrm the trend that the buried AGN fraction increases with infrared luminosity: 17 ± 15% for normal IRGs, +0 72 ± 13% for LIRGs, and 100−37% for ULIRGs in our sample. This suggests that the presence buried AGN is ubiquitous in U/LIRGs.

2. The ratio L3.3PAH/LIR decreases with infrared luminosity. This is opposed to the previous

reports that L3.3PAH/LIR have a constant value (Imanishi, 2002; Mouri et al., 1990). This decrease is possibly due to three eﬀects: (1) destruction of PAH molecules by X-rays emitted from the AGN, (2) PAH destruction by merger processes, and (3) increase of AGN contribution to the total infrared luminosity.

3. The energy contribution from the AGN torus-dust emission to the total infrared luminosity also increases with infrared luminosity, but it only reaches ∼7% in LIRGs and ∼20% even in ULIRGs. This suggests that majority of the total infrared luminosity in U/LIRGs originates from SB, not from AGN.

4. Combining the above results with the luminosity function of infrared galaxies derived (AGN) by Goto et al. (2010, 2011), we estimate the AGN luminosity density of ΩIR = 5 3 (AGN) 4 3 4.4 × 10 L /Mpc for LIRGs and ΩIR = 2.6 × 10 L /Mpc for ULIRGs in the local universe.

(AGN) 5. LSF versus Lbol luminosity plot for our sample shows that infrared galaxies with buried AGN are located above the luminosity correlation line derived from optically-selected Seyfert galaxies. We interpret that infrared galaxies could be an early phase of the evolutional track of AGN.

82 5

Conclusion and Outlook

5.1 Conclusion

AGN are powered by accretion of gas and dust onto a SMBH. Unified model of AGN (Antonucci, 1993) proposed that the observed difference of type-1 and type-2 AGN are well explained by the presence of dusty tori. While the success of the unified model of AGN, the dusty torus itself remained poorly understood. Based on my thesis, I summarize the conclusions on the improvement of the understanding of AGN tori.

5.1.1 Constraints of the torus models on the market

Since torus emission has a peak at MIR wavelengths, we have gathered MIR fluxes of hard X-ray selected Swift/BAT 9-month AGN catalog (Tueller et al., 2008), which is the least biased sample against absorption up to NH ∼ 24.5. Based on the cross-correlation with AKARI, IRAS, and WISE all-sky infrared survey catalogs, we find good correlation between their hard X-ray and MIR luminosities over three orders of magnitude for all types of AGN including type-1, type-2, and new-type AGN, suggesting that the MIR emission from the dusty tori are isotropic. This results suggest that clumpy dusty tori rather than homogeneous ones. In addition, we also find excess signals around 9 µm in “new type” AGN, suggesting that their host galaxies have strong starburst activities.

5.1.2 Clumpy torus model application to the observed data

The above research suggested that the clumpy torus is more preferable rather than smooth one. The next step we have compiled is to apply the torus model to the observed dust-torus SEDs, where high spatial resolution infrared observations of AGN are crucial to disentangle the emission from the host galaxy components. Applying Bayesclumpy to the torus SEDs

83 5. CONCLUSION AND OUTLOOK

enable us to derive the torus geometry such as the parameters of torus size, scale height, and covering factor. We show that local AGN have torus size with ∼2 pc and the torus scale height with ∼1 pc. In addition to the result above, we divided the sample into subgroups based on whether or not they are optically type-1, type-2 with observational hidden broad line regions sign (HBLR), and type-2 without any observational broad line region signs (NHBLR). We found

that the tori of Type-1 AGN have smaller σ, Y , NH, and CT than those of HBLR and NHBLR.

Moreover, the tori of NHBLR are thicker and therefore have higher CT than those of HBLR. These diﬀerences the torus properties of HBLR and NHBLR AGN would make it more diﬃcult to detect hidden BLR in NHBLR.

5.1.3 Obscured AGN search obtained from infrared observations

Many studies imply that the cosmic evolution of obscured AGN are somehow coupled to the star formation activity, which peaks at z ∼ 2 (Hopkins et al., 2006). Therefore, complete AGN surveys covering heavily obscured (=buried) AGN including Compton-thick populations are crucial to understand the cosmic history hidden by dust/gas obscuration. Candidates of galaxies that

host such buried AGN are infrared galaxies, deﬁned as those having infrared luminosity (LIR) 10 with LIR > 10 L . We conducted AKARI 2.5–5.0 µm spectroscopy of such infrared galaxies without any signs of AGN in the optical and X-ray bands. Applying our AGN diagnostics to the AKARI spectra by decomposing the spectra into stellar and AGN dust-torus components, we found that both the fraction and the energy contribution of buried AGN increase with infrared 10 13 luminosity from 10 L to 10 L . However, the energy contribution from AGN in the total infrared luminosity is only ∼7% in LIRGs and ∼20% in ULIRGs, suggesting that the majority of the infrared luminosity originates from starburst activity, not AGN. This simple method will have a great advantage in the era of JWST, because we can apply our method to high-z galaxies by observing the rest 2.5–5.0 µm band (5.0–10.0 µm at z = 1 and 7.5–15.0 µm at z = 2) with JWST /MIRI (5–28 µm). Other methods based on the Spitzer bandpass (5–35 µm) would have diﬃculty studying high-z objects; for instance, the observed spectral range corresponds to 15– 105 µm at z = 2, which only the SPICA mission can cover. Thus, we have essential method for studying infrared galaxies in the distant universe even before the launch of SPICA (from 2028–).

5.2 Future Prospects for AGN Studies through Dusty Tori

Thanks to the help of clumpy torus modeling to the data obtained from high spatial resolution MIR observations, we estimated the physical size of the torus (∼ 1 pc) and disclosed that both

84 5.2 Future Prospects for AGN Studies through Dusty Tori

SMBHs (< 0.001 pc) and nuclear star formation (∼ 100 pc) have strong relations between AGN tori (Imanishi et al., 2011a). However, the sample of previous studies was limited to 1) luminous AGN both in the X-ray and mid-infrared bands and moderately obscured AGN up to log NH < 24.5. Recent studies reported mysterious AGN that could have been in the diﬀerent evolutional stage and have not been discussed well because of the rare detections of those populations in the previous studies. Here, I mention these issues and show my future research plan on the application of the torus model to such AGN. I also plan to trace evolutional track from buried AGN to unobscured AGN using statistically signiﬁcant sample. Finally SMA and ALMA observations enable us to unveil the connection between AGN torus and nuclear starburst activity over wide range of luminosity 40 45 25 −2 (10 < LX < 10 erg/s) and torus obscuration (NH < 10 cm ). The projects I describe here will give us the complete view of diverse AGN 1) torus properties, 2) evolution, and 3) the origins that still remain uncertain. The high spatial resolution infrared/radio observations with Subaru and SMA/ALMA are crucial to achieve my future plan (Xu et al., 2014).

5.2.1 Unveiling the torus properties of Low-Luminosity AGN (LLAGN)

Target luminosity range in this study Luminosity range in the previous study

Figure 5.1: Obscured AGN fraction (=covering factor) as a function of AGN luminosity. - The original ﬁgure was taken from Burlon et al. (2011).

Many studies (Hasinger, 2008; Maiolino et al., 2007; Ricci et al., 2013; Simpson, 2005;

Ueda et al., 2003, 2014) showed that the covering factor (CT) decreases as a function of AGN luminosity. This suggests that higher AGN power sublimates the inner torus radius more eﬃciently, and the geometrical covering factor resulted in decrease (so called “receding torus

85 5. CONCLUSION AND OUTLOOK

model”; Lawrence, 1991). However, recent studies based on hard X-ray observations (e.g.,

Burlon et al. 2011) reported that the fraction of obscured AGN has a peak around LX ∼ 1042−43 erg/s below which it starts to decrease towards lower luminosity (See Figure 5.1), suggesting low-luminosity AGN (LLAGN) have different environment compared to luminous AGN which follow the simple receding torus model. To understand this phenomena in a unified way over the wide range of luminosity, it is crucial to establish the trend and reveal the origin. However, X-ray observations cannot remove uncertainty anymore because X-ray emission of LLAGN could be strongly contaminated from the starburst component in the host galaxies due to their intrinsic faintness and bad spatial resolution (∼arcmin). High spatial infrared observations are the most reasonable solutions to investigate the torus properties without any contamination from host galaxies. Considering the peak of the torus emission locates around 10–30 µm (e.g., Mullaney et al. 2011), compiling the SED from near- IR, mid-IR, and even far-IR is crucial to obtain more reliable torus parameters by taking into account the colder dust. “Los Piratas AGN project”, where I belong, has been conduct- 40−43 ing intensive infrared observations of nearby ∼ 20 LLAGN with LX ∼ 10 erg/s using GTC/CanariCam and already obtained several N (∼ 10 µm) band spectroscopy/imaging data. I also plan to observe Q (∼ 20 µm) band imaging to constrain the peak of torus emission of those LLAGN with Subaru/COMICS, which is the only instrument capable of Q band observations without any atmospheric difficulties. Furthermore, some of LLAGN could not be detected in the mid-IR band even with 8-m class telescopes due to their faintness. Thus, the observation of ALMA telescope is essential to cover far-IR (λ > 500 µm) torus SEDs, where only ALMA can achieve crucial spatial resolution (∼0.01 arcsec) in this band to isolate the torus emission. In a final stage, by applying Bayesclumpy to the torus SEDs (e.g., the same method in Chapter 3), torus properties including torus size and covering factor of ∼20 LLAGN will be obtained, which is a statistically significant number of sources to discuss the torus geometry. Combining my ongoing study with Ichikawa et al. 2015, whole AGN torus properties will be unveiled with 40 45 wide range of AGN luminosity with 10 < LX < 10 erg/s.

5.2.2 Origin of dust-free AGN (DFAGN) in the local universe

In the study of Chapter 2, we statistically examined the MIR properties of a complete sample of local AGN detected in the Swift/BAT all-sky hard X-ray survey. Out of 135 non-blazar AGN in the Swift/BAT 9-month catalog, we unambiguously identiﬁed the MIR photometric data of 128 sources. Interestingly, we also found AGN with very weak MIR emission (2 orders of magnitude smaller than the typical MIR luminosities considering their hard X-ray luminosity correlations, shown with the black triangle in the Figure 5.2), which were not detected even

86 5.2 Future Prospects for AGN Studies through Dusty Tori

45.5

44.5

m) [erg/s] 43.5 µ

(9 43 λ

L 42.5 λ 42 log 41.5

41 41 41.5 42 42.5 43 43.5 44 44.5 45 45.5 log LHX(14-195 keV) [erg/s]

Figure 5.2: Luminosity correlation between 9 µm and hard X-ray band. - Blue square and red circle represents type-1 and type-2 AGN, respectively. Black triangle represents 9 µm undetected source with the upper limit of 9 µm luminosity. by the high sensitivity limit of WISE. MIR emission of AGN dominantly originates from dust- torus. Although these sources have enough X-ray emission to heat the dust, to explain the weak MIR emission, they should have very small, or possibly even absent dusty tori (hereafter we call them “dust-free AGN” (DFAGN)). Considering DFAGN are luminous in X-rays but faint in the MIR band, this situation suggests DFAGN are in a final stage of AGN consuming final fuel from the the torus accretion, and we can finally investigate how the AGN quench. Investigation of these AGN are crucial. One important first step is to make a catalog of DFAGN. We already discovered 7 DFAGN from Swift/BAT 9-month catalog. Currently, Swift/BAT 70-month catalog are now available and in which over 700 AGN are registered. We already tentatively found 30 DFAGN candidates out of 700 sources. Although these sources are too faint in MIR band even with 8-m class telescopes such as Subaru, in a near future in the era of TMT, there is a possibility to obtain the MIR image of such DFAGN.

5.2.3 ALMA observations of those AGN populations

It is still unclear how such various types of low-luminosity (Chapter 5.2.1), dust-free (Chapter 5.2.2), and buried AGN (Chapter 4) can be produced. LLAGN and DFAGN have small or lack of dusty tori (=small torus covering factor), and buried AGN have large tori (=large torus covering factor). One interesting possibility is that signiﬁcant star forming activity is taking

87 5. CONCLUSION AND OUTLOOK

place in the nuclear region, which gives sufficient thermal pressure to sustain the torus structure as discussed in Chapter 3. Indeed, recent numerical simulations support this picture (Wada & Norman 2002). To confirm this scenario, it is crucial to reveal the presence/absence of active starbursts in the nuclear region, which can be traced by molecular gases. SMA and ALMA provides us with an ideal opportunity by its unprecedented sensitivities for molecular lines. Future CO(2-1) observations for LL-/DF-/buried AGN over wide range of AGN luminosity 40 45 25 −2 (10 < LX < 10 erg/s) without any obscuration biases (NH < 10 cm ) will give us the key insights of AGN evolutions of those populations. Immediate objective is measuring the gravitational stability of nuclear molecular gas to check whether they are unstable enough to sustain star-formation in the central region. It enable us to obtain the molecular gas mass Mgas via CO-toH2 conversion factor, and to see the velocity field of the central 100 pc region. With the observed velocity field and CO line profile, dynamical mass Mdyn will be derived. This enable us to access the gravitational stability (or instability) of the central molecular gas disk −1 in these nuclei to measure Toomore’s Q parameter because Q ∝ (Mgas/Mdyn) . According to Sakamoto et al. (1999), nearby starburst galaxies have (Mgas/Mdyn) > 0.1 in the central region. Therefore if such buried AGN exceed (Mgas/Mdyn) = 0.1 significantly, the molecular gas disk can be unstable. And, low-luminosity/dust-free AGN would be expected to have lower

(Mgas/Mdyn) with less than 0.1. These results will naturally support for our working hypothesis that covering factor of AGN are highly affected by massive starburst, which may result in a geometrically thick torus structure. Furthermore, 1.2 mm continuum emission will also come for free because the expected torus emission based on the extrapolation from the torus model fitting should be sufficient to detect. Therefore, this observation can obtain the torus SEDs at 1.2 mm simultaneously and it will strongly constrain the torus properties as discussed in Chapter 3. To this end, with reasonable and realistic steps, first proposing to use SMA to observe CO(2- 1) for bright and nearby LL-/DF-/buried AGN in order to check this method is achievable for distant sources with ALMA. Then proposing ALMA CO(2-1) observations in Band 6 for the remaining fainter and distant AGN to obtain the complete view of nuclear activity of various AGN.

88 Awaki, H., Anabuki, N., Fukazawa, Y., et al. 2008, PASJ, 60, 293 27

Baade, W., & Minkowski, R. 1954, ApJ, 119, 206 2

Ballantyne, D. R. 2005, MNRAS, 362, 1183 27 References Baumgartner, W. H., Tueller, J., Markwardt, C. B., et al. 2013, ApJS, 207, 19

Bauschlicher, C. W., Jr., Peeters, E., & Allaman- dola, L. J. 2009, ApJ, 697, 311 Ajello, M., Greiner, J., Sato, G., et al. 2008, ApJ, 689, 666 15 Becker, R. H., Helfand, D. J., White, R. L., Gregg, M. D., & Laurent-Muehleisen, S. A. Alexander, D. M., Hough, J. H., Young, S., et al. 2003, VizieR Online Data Catalog, 8071, 0 1999, MNRAS, 303, L17 43 Beckmann, V., Soldi, S., Ricci, C., et al. 2009, Alexander, D. M. 2001, MNRAS, 320, L15 11, A&A, 505, 417 13 41, 60 Beichman, C. A., Neugebauer, G., Habing, H. J., Alonso-Herrero, A., Quillen, A. C., Rieke, G. H., Clegg, P. E., & Chester, T. J. 1988, Infrared Ivanov, V. D., & Efstathiou, A. 2003, AJ, 126, astronomical satellite (IRAS) catalogs and at- 81 42, 48 lases. Volume 1: Explanatory supplement, 1, 22, 60 Alonso-Herrero, A., Rieke, G. H., Rieke, M. J., et al. 2006, ApJ, 650, 835 Bianchi, S., Piconcelli, E., Chiaberge, M., et al. 2009, ApJ, 695, 781 27 Alonso-Herrero, A., Ramos Almeida, C., Mason, R., et al. 2011, ApJ, 736, 82 11, 42, 43, 44, 46, Bird, A. J., Bazzano, A., Bassani, L., et al. 2010, 47, 48, 49, 50, 53, 56, 60, 68 ApJS, 186, 1 20

Alonso-Herrero, A., S´anchez-Portal, M., Ramos Blanco, P. R., Ward, M. J., & Wright, G. S. 1990, Almeida, C., et al. 2012, MNRAS, 425, 311 MNRAS, 242, 4P 11, 41 16, 43, 45, 78 Bolton, J. G., & Stanley, G. J. 1948, Nature, 161, Alonso-Herrero, A., Pereira-Santaella, M., Rieke, 312 1 G. H., & Rigopoulou, D. 2012, ApJ, 744, 2 45 Bolton, J. G. 1948, Nature, 162, 141 2

Antonucci, R. R. J., & Cohen, R. D. 1983, ApJ, Brandt, W. N., & Hasinger, G. 2005, ARA&A, 271, 564 9 43, 827 19

Antonucci, R. R. J., & Miller, J. S. 1985, ApJ, Brewer, B. J., Elliott, T. M. 2014, MNRAS, 439, 297, 621 5, 11, 41 L31 54

Antonucci, R. 1993, ARA&A, 31, 473 ii, 5, 41, Brightman, M., & Nandra, K. 2011, MNRAS, 83 413, 1206 15, 43, 63

Aretxaga, I., Joguet, B., Kunth, D., Melnick, J., Brightman, M., & Ueda, Y. 2012, MNRAS, 423, & Terlevich, R. J. 1999, ApJ, 519, L123 8 702 14, 19

Asensio Ramos, A., & Ramos Almeida, C. 2009, Brightman, M., Nandra, K., Salvato, M., et al. ApJ, 696, 2075 11, 16, 47, 50 2014, MNRAS, 443, 1999 19

Asmus, D., Gandhi, P., Smette, A., H¨onig,S. F., Bryant, J. J., & Hunstead, R. W. 1999, MNRAS, & Duschl, W. J. 2011, A&A, 536, AA36 20, 42 308, 431

89 REFERENCES

Burlon, D., Ajello, M., Greiner, J., et al. 2011, Elitzur, M., Ho, L. C., & Trump, J. R. 2014, ApJ, 728, 58 13, 27, 85 MNRAS, 438, 3340 47

Burtscher, L., Jaﬀe, W., Raban, D., et al. 2009, Elitzur, M. 2008, New Astronomy Reviews, 52, ApJ, 705, L53 8 274 12

Capetti, A., Axon, D. J., Macchetto, F., Sparks, Elitzur, M. 2012, ApJ, 747, L33 57 W. B., & Boksenberg, A. 1996, ApJ, 466, 169 Elvis, M., Wilkes, B. J., McDowell, J. C., et al. Carico, D. P., Sanders, D. B., Soifer, B. T., et al. 1994, ApJS, 95, 1 43 1988, AJ, 95, 356 65, 81 Emmering, R. T., Blandford, R. D., & Shlosman, Carollo, C. M., Stiavelli, M., Seigar, M., de I. 1992, ApJ, 385, 460 12 Zeeuw, P. T., & Dejonghe, H. 2002, AJ, 123, 159 46 Esquej, P., Alonso-Herrero, A., Gonz´alez-Mart´ın, O., et al. 2014, ApJ, 780, 86 60 Castro, A., Miyaji, T., Shirahata, M., et al. 2014, arXiv:1408.3172 60 Everett, J. E. 2005, ApJ, 631, 689 12

Condon, J. J., Helou, G., Sanders, D. B., & Fath, E. A. 1909, Lick Observatory Bulletin, 5, Soifer, B. T. 1996, ApJS, 103, 81 71 1

Condon, J. J., Cotton, W. D., Greisen, E. W., et Feltre, A., Hatziminaoglou, E., Fritz, J., & al. 1998, AJ, 115, 1693 Franceschini, A. 2012, MNRAS, 426, 120 42

Deluit, S. J. 2004, A&A, 415, 39 Galliano, E., & Alloin, D. 2008, A&A, 487, 519 46 Deo, R. P., Richards, G. T., Crenshaw, D. M., & Kraemer, S. B. 2009, ApJ, 705, 14 45, 46 Gandhi, P., Horst, H., Smette, A., et al. 2009, A&A, 502, 457 16, 20, 33, 34, 35, 36, 39, 42, Di Matteo, T., Springel, V., & Hernquist, L. 80 2005, Nature, 433, 604 77 Garc´ıa-Burillo, S., Combes, F., Usero, A., et al. Draine, B. T. 1989, Infrared Spectroscopy in As- 2014, A&A, 567, A125 55 tronomy, 290, 93 70 Gonz´alez-Mart´ın, O., Papadakis, I., Braito, V., D´ıaz-Santos, T., Alonso-Herrero, A., Colina, L., et al. 2011, A&A, 527, AA142 27 et al. 2008, ApJ, 685, 211 Gonz´alez-Mart´ın,O., Rodr´ıguez-Espinosa, J. M., D´ıaz-Santos, T., Alonso-Herrero, A., Colina, L., D´ıaz-Santos, T., et al. 2013, A&A, 553, A35 et al. 2010, ApJ, 711, 328 43, 44, 48, 50, 60

Edelson, R. A., & Malkan, M. A. 1986, ApJ, 308, Goto, T., Takagi, T., Matsuhara, H., et al. 2010, 59 6 A&A, 514, A6 16, 64, 79, 82

Efstathiou, A., & Rowan-Robinson, M. 1995, Goto, T., Arnouts, S., Inami, H., et al. 2011, MN- MNRAS, 273, 649 6, 42 RAS, 410, 573 79, 82

Eguchi, S., Ueda, Y., Terashima, Y., Mushotzky, Goulding, A. D., Alexander, D. M., Bauer, F. E., R., & Tueller, J. 2009, ApJ, 696, 1657 15, 20, et al. 2012, ApJ, 755, 5 14, 39, 43, 48, 60 27 Granato, G. L., Danese, L., & Franceschini, A. Eguchi, S., Ueda, Y., Awaki, H., et al. 2011, ApJ, 1997, ApJ, 486, 147 729, 31 15, 27 Greenstein, J. L., & Schmidt, M. 1964, ApJ, 140, Elitzur, M., & Ho, L. C. 2009, ApJ, 701, L91 12 1 2

Elitzur, M., & Shlosman, I. 2006, ApJ, 648, L101 Gu, Q., Maiolino, R., & Dultzin-Hacyan, D. 2001, 12 A&A, 366, 765 12, 42, 43

90 REFERENCES

Guainazzi, M., Rodriguez-Pascual, P., Fabian, Ichikawa, K., Imanishi, M., Ueda, Y., et al. 2014, A. C., Iwasawa, K., & Matt, G. 2004, MN- ApJ, 794, 139 60 RAS, 355, 297 43 Ichikawa, K., et al. 2014, submitted to ApJ Hao, L., Spoon, H. W. W., Sloan, G. C., et al. 2005, ApJ, 625, L75 7 Ikeda, S., Awaki, H., & Terashima, Y. 2009, ApJ, 692, 608 15, 63 Hao, L., Weedman, D. W., Spoon, H. W. W., et al. 2007, ApJ, 655, L77 7, 37 Imanishi, M., & Dudley, C. C. 2000, ApJ, 545, 701 14, 37, 64, 67 Hasinger, G. 2008, A&A, 490, 905 12, 33, 85 Imanishi, M., & Maloney, P. R. 2003, ApJ, 588, Heckman, T. M., Armus, L., & Miley, G. K. 1990, 165 70, 71 ApJS, 74, 833 Imanishi, M., Dudley, C. C., & Maloney, P. R. Heisler, C. A., Lumsden, S. L., & Bailey, J. A. 2006, ApJ, 637, 114 64 1997, Nature, 385, 700 11, 42, 47, 59 Imanishi, M., Nakagawa, T., Ohyama, Y., et al. Ho, L. C., & Peng, C. Y. 2001, ApJ, 555, 650 2008, PASJ, 60, 489 64, 65, 68, 69

Ho, L. C., Filippenko, A. V., & Sargent, W. L. W. Imanishi, M., Nakagawa, T., Shirahata, M., 1997, ApJS, 112, 315 44 Ohyama, Y., & Onaka, T. 2010, ApJ, 721, 1233 64, 65, 66, 68, 69, 70, 71, 76 Ho, L. C. 2002, ApJ, 564, 120 Imanishi, M., Ichikawa, K., Takeuchi, T., et al. Hopkins, A. M., & Beacom, J. F. 2006, ApJ, 651, 2011, PASJ, 63, 447 14, 60, 80, 85 142 14, 63 Imanishi, M., Imase, K., Oi, N., & Ichikawa, K. Hopkins, P. F., Hernquist, L., Cox, T. J., et al. 2011, AJ, 141, 156 64 2006, ApJS, 163, 1 77, 84 Imanishi, M. 2002, ApJ, 569, 44 43, 82 Horst, H., Gandhi, P., Smette, A., & Duschl, Ishihara, D., Onaka, T., Kataza, H., et al. 2010, W. J. 2008, A&A, 479, 389 20 A&A, 514, AA1 23

Horst, H., Duschl, W. J., Gandhi, P., & Smette, Itoh, T., Done, C., Makishima, K., et al. 2008, A. 2009, A&A, 495, 137 PASJ, 60, 251 27, 43

Hunt, L. K., Giovanardi, C., & Helou, G. 2002, Jaﬀe, W., Meisenheimer, K., R´’ottgering, A&A, 394, 873 69, 70 H. J. A., et al. 2004, Nature, 429, 47 8

Hönig,S. F., & Kishimoto, M. 2010, A&A, 523, Jansky, K. G. 1933, Nature, 132, 66 1 A27 9, 10, 42 Jiménez-Bailón, E., Santos-Lleó, M., Dahlem, Hönig,S. F., Beckert, T., Ohnaka, K., & Weigelt, M., et al. 2005, A&A, 442, 861 44 G. 2006, A&A, 452, 459 9, 10, 36, 42 Kawada, M., Baba, H., Barthel, P. D., et al. 2007, Hönig,S. F., Leipski, C., Antonucci, R., & Haas, PASJ, 59, 389 21 M. 2011, ApJ, 736, 26 42 Kawaguchi, T., & Mori, M. 2010, ApJ, 724, L183 Hönig,S. F., Gandhi, P., Asmus, D., et al. 2014, 10 MNRAS, 438, 647 45 Kawaguchi, T., & Mori, M. 2011, ApJ, 737, 105 Hubble, E. P. 1926, ApJ, 64, 321 1 10, 53

Ichikawa, K., Ueda, Y., Terashima, Y., et al. Kawamuro, T., Ueda, Y., Tazaki, F., & 2012, ApJ, 754, 45 16, 42, 60 Terashima, Y. 2013, ApJ, 770, 157 13

Ichikawa, K., Ueda, Y., Terashima, Y., et al. Kellermann, K. I., Sramek, R., Schmidt, M., 2012, Torus Workshop, 2012, 109 12, 60 Shaﬀer, D. B., & Green, R. 1989, AJ, 98, 1195

91 REFERENCES

Khachikian, E. Y., & Weedman, D. W. 1971, Lindblad, P. O. 1999, A&A Rev., 9, 221 ApJ, 164, L109 9 Lira, P., Videla, L., Wu, Y., et al. 2013, ApJ, 764, Kishimoto, M., H´’onig, S. F., Beckert, T., & 159 42, 68 Weigelt, G. 2007, A&A, 476, 713 53 Lopez-Rodriguez, E., Packham, C., Young, S., et Kishimoto, M., H´’onig, S. F., Antonucci, R., et al. 2013, MNRAS, 431, 2723 13 al. 2009, A&A, 507, L57 53 Lu, N., Helou, G., Werner, M. W., et al. 2003, Kishimoto, M., H´’onig, S. F., Antonucci, R., et ApJ, 588, 199 69 al. 2011, A&A, 536, 78 6, 7, 53, 54, 68 Lumsden, S. L., Heisler, C. A., Bailey, J. A., Klaas, U., Haas, M., M´’uller, S. A. H., et al. 2001, Hough, J. H., & Young, S. 2001, MNRAS, 327, A&A, 379, 823 65, 81 459 47, 59

Konigl, A., & Kartje, J. F. 1994, ApJ, 434, 446 Lumsden, S. L., Alexander, D. M., & Hough, 12 J. H. 2004, MNRAS, 348, 1451 12, 42

Koshida, S., Yoshii, Y., Kobayashi, Y., et al. Lutz, D., Feuchtgruber, H., Genzel, R., et al. 2009, ApJ, 700, L109 53 1996, A&A, 315, L269 64

Koss, M., Mushotzky, R., Veilleux, S., et al. 2011, Lynden-Bell, D. 1969, Nature, 223, 690 2, 3 ApJ, 739, 57 21 Magnelli, B., Elbaz, D., Chary, R. R., et al. 2011, Krabbe, A., B¨oker, T., & Maiolino, R. 2001, ApJ, A&A, 528, A35 16, 64 557, 626 20 Magorrian, J., Tremaine, S., Richstone, D., et al. Kristian, J. 1973, ApJ, 179, L61 2 1998, AJ, 115, 228 13

Krolik, J. H., & Begelman, M. C. 1986, ApJ, 308, Maiolino, R., & Rieke, G. H. 1995, ApJ, 454, 95 L55 5, 63 44

Krolik, J. H., & Begelman, M. C. 1988, ApJ, 329, Maiolino, R., Comastri, A., Gilli, R., et al. 2003, 702 6 MNRAS, 344, L59 16 Maiolino, R., Shemmer, O., Imanishi, M., et al. Kullback, S., & Leibler, A. 1951, Ann. Math. 2007, A&A, 468, 979 12, 33, 85 Stat., 22, 79 56 Malkan, M. A., Gorjian, V., & Tam, R. 1998, Landt, H., Ward, M. J., Elvis, M., & Karovska, ApJS, 117, 25 37 M. 2014, arXiv:1401.1040 Marchese, E., Della Ceca, R., Caccianiga, A., et Lawrence, A. 1991, MNRAS, 252, 586 11, 12, 57, al. 2012, A&A, 539, A48 34, 43, 80 86 Marin, F. 2014, MNRAS, 441, 551 47 Le Floc’h, E., Papovich, C., Dole, H., et al. 2005, ApJ, 632, 169 16, 64 Marinucci, A., Bianchi, S., Nicastro, F., Matt, G., 34, 43, 80 Lee, J. C., Hwang, H. S., Lee, M. G., Kim, M., & Lee, J. H. 2012, ApJ, 756, 95 78 Mason, R. E., Lopez-Rodriguez, E., Packham, C., et al. 2012, AJ, 144, 11 13 Levenson, N. A., Weaver, K. A., Heckman, T. M., Awaki, H., & Terashima, Y. 2004, ApJ, 602, Mason, R. E., Ramos Almeida, C., Levenson, 135 N. A., Nemmen, R., & Alonso-Herrero, A. 2013, ApJ, 777, 164 13 Levenson, N. A., Heckman, T. M., Krolik, J. H., Weaver, K. A., &Zdotycki, P. T. 2006, ApJ, Matsuta, K., Gandhi, P., Dotani, T., et al. 2012, 648, 111 43, 50 ApJ, 753, 104 36

Levenson, N. A., Radomski, J. T., Packham, C., Meisenheimer, K., Tristram, K. R. W., Jaﬀe, W., et al. 2009, ApJ, 703, 390 8, 10, 20 et al. 2007, A&A, 471, 453 46

92 REFERENCES

Meléndez, M., Kraemer, S. B., Armentrout, Nenkova, M., Sirocky, M. M., Ivezić, Z.,ˇ & B. K., et al. 2008, ApJ, 682, 94 36 Elitzur, M. 2008, ApJ, 685, 147 9, 11, 16, 36, 42, 47, 49, 50, 68 Meléndez,M., Kraemer, S. B., Schmitt, H. R., et al. 2008, ApJ, 689, 95 36 Nenkova, M., Sirocky, M. M., Nikutta, R., Ivezić, Z.,ˇ & Elitzur, M. 2008, ApJ, 685, 160 9, 11, Meléndez,M., Mushotzky, R. F., Shimizu, T. T., 16, 36, 42, 49, 68 Barger, A. J., & Cowie, L. L. 2014, ApJ, 794, 152 46 Nenkova, M., Sirocky, M. M., Nikutta, R., Ivezić, Z.,ˇ & Elitzur, M. 2010, ApJ, 723, 1827 Mihos, J. C., & Hernquist, L. 1996, ApJ, 464, 641 77 Netzer, H., & Laor, A. 1993, ApJ, 404, L51

Minkowski, R. 1960, ApJ, 132, 908 2 Netzer, H., Lira, P., Trakhtenbrot, B., Shemmer, O., & Cury, I. 2007, ApJ, 671, 1256 45 Moorwood, A. F. M. 1986, A&A, 166, 4 64 Netzer, H. 2009, MNRAS, 399, 1907 80, 81 Mor, R., & Netzer, H. 2012, MNRAS, 420, 526 Neugebauer, G., Habing, H. J., van Duinen, R., Mor, R., & Trakhtenbrot, B. 2011, ApJ, 737, L36 et al. 1984, ApJ, 278, L1 20 Mor, R., Netzer, H., & Elitzur, M. 2009, ApJ, Nicastro, F., Martocchia, A., & Matt, G. 2003, 705, 298 ApJ, 589, L13 43 Moran, E. C., Barth, A. J., Kay, L. E., & Filip- penko, A. V. 2000, ApJ, 540, L73 43, 47 Nikutta, R., Elitzur, M., & Lacy, M. 2009, ApJ, 707, 1550 42 Moran, E. C., Filippenko, A. V., & Chornock, R. 2002, ApJ, 579, L71 47 Nishiyama, S., Tamura, M., Hatano, H., et al. 2009, ApJ, 696, 1407 64 Mouri, H., Kawara, K., Taniguchi, Y., & Nishida, M. 1990, ApJ, 356, L39 67, 82 Noguchi, K., Terashima, Y., Ishino, Y., et al. 2010, ApJ, 711, 144 27 Mullaney, J. R., Alexander, D. M., Goulding, A. D., & Hickox, R. C. 2011, MNRAS, 414, Ohyama, Y., Onaka, T., Matsuhara, H., et al. 1082 78 2007, PASJ, 59, 411 66

Murakami, H., Baba, H., Barthel, P., et al. 2007, Oi, N., Imanishi, M., & Imase, K. 2010, PASJ, PASJ, 59, 369 20, 65, 66 62, 1509 80

Murphy, E. J., Chary, R.-R., Dickinson, M., et Oke, J. B., & Sargent, W. L. W. 1968, ApJ, 151, al. 2011, ApJ, 732, 126 16, 64 807 2

Mushotzky, R. F., Winter, L. M., McIntosh, Onaka, T., Matsuhara, H., Wada, T., et al. 2007, D. H., & Tueller, J. 2008, ApJ, 684, L65 33 PASJ, 59, 401 21, 65, 66

Nagar, N. M., Oliva, E., Marconi, A., Maiolino, Ossenkopf, V., Henning, T., & Mathis, J. S. 1992, R. 2002, A&A, 391, L21 11, 41 A&A, 261, 567 48

Nardini, E., Risaliti, G., Salvati, M., et al. 2008, Oyabu, S., Ishihara, D., Malkan, M., et al. 2011, MNRAS, 385, L130 16 A&A, 529, A122 19, 69

Nardini, E., Risaliti, G., Salvati, M., et al. 2009, Packham, C., Radomski, J. T., Roche, P. F., et MNRAS, 399, 1373 16 al. 2005, ApJ, 618, L17 8

Nardini, E., Risaliti, G., Watabe, Y., Salvati, M., Peng, Z., Gu, Q., Melnick, J., & Zhao, Y. 2006, & Sani, E. 2010, MNRAS, 405, 2505 16, 78 A&A, 453, 863 45, 46

Nenkova, M., Ivezi´c, Z.,ˇ & Elitzur, M. 2002, ApJ, Peterson, B. M., Ferrarese, L., Gilbert, K. M., et 570, L9 9, 10, 42 al. 2004, ApJ, 613, 682

93 REFERENCES

Petric, A. O., Armus, L., Howell, J., et al. 2011, Risaliti, G., Maiolino, R., Marconi, A., et al. ApJ, 730, 28 78 2006, MNRAS, 365, 303 64

Phillips, A. C. 1993, AJ, 105, 486 Risaliti, G., Braito, V., Laparola, V., et al. 2009, ApJ, 705, L1 27 Pier, E. A., & Krolik, J. H. 1992, ApJ, 401, 99 6, 7, 36, 42 Risaliti, G., Imanishi, M., & Sani, E. 2010, MN- RAS, 401, 197 64, 70, 71, 78 Pier, E. A., & Krolik, J. H. 1993, ApJ, 418, 673 Rowan-Robinson, M. 1977, ApJ, 213, 635 5 6, 7, 36, 42 Rush, B., Malkan, M. A., & Spinoglio, L. 1993, Polletta, M., Tajer, M., Maraschi, L., et al. 2007, ApJS, 89, 1 66 ApJ, 663, 81 37 Sandage, A. 1965, ApJ, 141, 1560 2 Quillen, A. C., McDonald, C., Alonso-Herrero, A., et al. 2001, ApJ, 547, 129 46 Sanders, D., & Ishida, C. 2004, The Neutral ISM in Starburst Galaxies, 320, 230 77 Raban, D., Jaffe, W., R´’ottgering, H., Meisen- heimer, K., & Tristram, K. R. W. 2009, MN- Sanders, D. B., & Mirabel, I. F. 1996, ARA&A, RAS, 394, 1325 8 34, 749 15, 68, 69 Sanders, D. B., Soifer, B. T., Elias, J. H., et al. Radomski, J. T., Packham, C., Levenson, N. A., 1988, ApJ, 325, 74 15, 77 et al. 2008, ApJ, 681, 141 8 Sani, E., Risaliti, G., Salvati, M., et al. 2008, Ramos Almeida, C., PérezGarc´ıa,A. M., Acosta- ApJ, 675, 96 64 Pulido, J. A., González-Mart´ın,O. 2008, ApJ, 680, L17 11, 41 Sargsyan, L., Weedman, D., Lebouteiller, V., et al. 2011, ApJ, 730, 19 38 Ramos Almeida, C., Levenson, N. A., Rodr´ıguez Espinosa, J. M., et al. 2009, ApJ, 702, 1127 Sawicki, M. 2002, AJ, 124, 3050 69 11, 42, 43, 46, 68 Scarrott, S. M., Draper, P. W., Stockdale, D. P., Ramos Almeida, C., Levenson, N. A., Alonso- & Wolstencroft, R. D. 1993, MNRAS, 264, L7 Herrero, A., et al. 2011, ApJ, 731, 92 10, 11, 44 42, 56, 57, 68 Schartmann, M., Meisenheimer, K., Camenzind, M., et al. 2008, A&A, 482, 67 9, 10, 36, 42 Ramos Almeida, C., Sánchez-Portal, M., Pérez Garc´ıa, A. M., et al. 2011, MNRAS, 417, L46 Schmitt, H. R. 2001, AJ, 122, 2243 12, 54 55 Seyfert, C. K. 1943, ApJ, 97, 28 1 Ramos Almeida, C., Alonso-Herrero, A., Leven- Shakura, N. I., & Sunyaev, R. A. 1973, A&A, 24, son, N. A., et al. 2014, MNRAS, 439, 3847 44, 337 3 50, 56 Siebenmorgen, R. 1993, ApJ, 408, 218 69 Reunanen, J., Kotilainen, J. K., & Prieto, M. A. 2003, MNRAS, 343, 192 11, 41 Simpson, C., Wilson, A. S., Bower, G., et al. 1997, ApJ, 474, 121 Reunanen, J., Prieto, M. A., & Siebenmorgen, R. 2010, MNRAS, 402, 879 Simpson, C. 1998, ApJ, 509, 653 46

Ricci, C., Walter, R., Courvoisier, T. J.-L., & Simpson, C. 2005, MNRAS, 360, 565 12, 33, 85 Paltani, S. 2011, A&A, 532, A102 57 Slipher, V. M. 1917, Proceedings of the American Philosophical Society, 56, 403 1 Ricci, C., Paltani, S., Awaki, H., et al. 2013, A&A, 553, A29 33, 57, 85 Smith, F. G. 1951, Nature, 168, 555 2 Ricci, C., Ueda, Y., Ichikawa, K., et al. 2014, Soifer, B. T., Sanders, D. B., Madore, B. F., et A&A, 567, A142 57 al. 1987, ApJ, 320, 238 81

94 REFERENCES

Soifer, B. T., Neugebauer, G., Matthews, K., et Ueda, Y., Akiyama, M., Hasinger, G., Miyaji, T., al. 2001, AJ, 122, 1213 64 & Watson, M. G. 2014, ApJ, 786, 104 vii, 14, 15, 33, 63, 85 Spinoglio, L., Malkan, M. A., Rush, B., Carrasco, L., & Recillas-Cruz, E. 1995, ApJ, 453, 616 69 Urry, C. M., & Padovani, P. 1995, PASP, 107, 803 6 Spinoglio, L., Andreani, P., & Malkan, M. A. 2002, ApJ, 572, 105 66, 81 Vasudevan, R. V., Fabian, A. C., Gandhi, P., Winter, L. M., & Mushotzky, R. F. 2010, MN- Stalevski, M., Fritz, J., Baes, M., Nakos, T., & RAS, 402, 1081 33 Popović, L. C.ˇ 2012, MNRAS, 420, 2756 10, 42, 53, 68 Veilleux, S., Kim, D.-C., Sanders, D. B., Maz- zarella, J. M., & Soifer, B. T. 1995, ApJS, 98, Storchi-Bergmann, T., Wilson, A. S., & Baldwin, 171 16 J. A. 1992, ApJ, 396, 45 Veilleux, S., Kim, D.-C., & Sanders, D. B. 1999, Storchi-Bergmann, T., Baldwin, J. A., & Wilson, ApJ, 522, 113 16 A. S. 1993, ApJ, 410, L11 9 Veilleux, S., Kim, D.-C., & Sanders, D. B. 2002, Suganuma, M., Yoshii, Y., Kobayashi, Y., et al. ApJS, 143, 315 76 2006, ApJ, 639, 4 6 Veilleux, S., Rupke, D. S. N., Kim, D.-C., et al. Tazaki, F., Ueda, Y., Terashima, Y., & 2009, ApJS, 182, 628 78 Mushotzky, R. F. 2011, ApJ, 738, 70 27 Videla, L., Lira, P., Andrews, H., et al. 2013, Terashima, Y., & Wilson, A. S. 2003, ApJ, 583, ApJS, 204, 23 68 145 Véron-Cetty, M.-P., & Véron,P. 2006, A&A, 455, Tielens, A. G. G. M. 2008, ARA&A, 46, 289 37 773 43

Tran, H. D., Lyke, J. E., & Mader, J. A. 2011, Vestergaard, M., & Peterson, B. M. 2006, ApJ, ApJ, 726, L21 11, 41 641, 689 3

Tran, H. D. 2001, ApJ, 554, L19 11, 41, 43, 47 Wada, K., & Norman, C. A. 2002, ApJ, 566, L21 60 Tran, H. D. 2003, ApJ, 583, 632 11, 41, 43, 47 Walker, M. F. 1968, ApJ, 151, 71 2 Tran, H. D. 2010, ApJ, 711, 1174 Wang, J.-M., & Zhang, E.-P. 2007, ApJ, 660, Tristram, K. R. W., Meisenheimer, K., Jaﬀe, W., 1072 43 et al. 2007, A&A, 474, 837 8 Weaver, K. A., Mel´endez, M., Mushotzky, R. F., Tristram, K. R. W., Raban, D., Meisenheimer, et al. 2010, ApJ, 716, 1151 36, 38 K., et al. 2009, A&A, 502, 67 8 Wilson, A. S., Shopbell, P. L., Simpson, C., et al. Tristram, K. R. W., & Schartmann, M. 2011, 2000, AJ, 120, 1325 5 A&A, 531, A99 8 Winkler, C., Courvoisier, T. J.-L., Di Cocco, G., et al. 2003, A&A, 411, L1 20 Tueller, J., Mushotzky, R. F., Barthelmy, S., et al. 2008, ApJ, 681, 113 19, 20, 21, 27, 31, 39, Winter, L. M., & Mushotzky, R. 2010, ApJ, 719, 43, 83 737 27 Turner, T. J., Miller, L., Kraemer, S. B., Reeves, Winter, L. M., Mushotzky, R. F., Reynolds, C. S., J. N., & Pounds, K. A. 2009, ApJ, 698, 99 27 & Tueller, J. 2009, ApJ, 690, 1322 15, 20, 21, 27 Ueda, Y., Akiyama, M., Ohta, K., & Miyaji, T. 2003, ApJ, 598, 886 12, 27, 33, 85 Winter, L. M., Mushotzky, R. F., Terashima, Y., & Ueda, Y. 2009, ApJ, 701, 1644 27 Ueda, Y., Eguchi, S., Terashima, Y., et al. 2007, ApJ, 664, L79 15, 20, 24 Woltjer, L. 1959, ApJ, 130, 38 2

95 REFERENCES

Woo, J.-H., Schulze, A., Park, D., et al. 2013, ApJ, 772, 49 13

Xu, C. K., Cao, C., Lu, N., et al. 2014, arXiv:1411.1111 85

Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868 20, 22

Yamada, R., Oyabu, S., Kaneda, H., et al. 2013, PASJ accepted, arXiv:1307.6356 69, 77

Yamamura, I., Makiuti, S., Ikeda, N., et al. 2010, VizieR Online Data Catalog, 2298, 0 23

Yuan, T.-T., Kewley, L. J., & Sanders, D. B. 2010, ApJ, 709, 884 16

Zdziarski, A. A., Lubi´nski, P., & Smith, D. A. 1999, MNRAS, 303, L11 4 999

96 REFERENCES