Modern Quasar SEDs Zhaohui Shang Tianjin Normal University

Modern Quasar SEDs Zhaohui Shang （Tianjin Normal University） Kunming, Feb. 2009

Quasar Spectral Energy Distributions (SED) • Significant energy output over wide frequency range • “Big blue bump” (UV bump) – strongest energy output • Infrared bump – energy output comparable to UV bump Quasar SED (Elvis et al. 1994) • Infrared broad band photometry

AGN Structure - Multi-wavelength Study Major components: • super massive black hole • accretion disk (opt. , UV, X-ray, continuum) • emision line clouds/wind • dusty torus (IR) • Jets (radio)

Example: Big Blue Bump – Spectral break at ~1000 Å (HST composite) HST AGN composite spectra Zheng et al. (1997) Telfer et al. (2002) • 101 objects • 284 HST spectra • z > 0. 33 • 184 objects • 332 HST spectra • z > 0. 33

Recent Results from Spitzer (broad band – IRAC) • 259 SDSS quasars (Richards et al. 2006) • Overall SEDs consistent with the mean SEDs of Elvis et al. 1994 • Large SED diversity for individual objects

Recent Results from Spitzer (broad band – IRAC, MIPS) • 13 high-redshift (z>4. 5) quasars (Hines et al. 2006, Ap. J, 641, L 85) • Consistent with SEDs of lowredshift quasars (Elvis et al. 1994)

Recent Results from Spitzer (IRS spectra) E. g, • 29 quasars (Netzer et al. 2007)

Importance of Quasar SEDs, practically • Important in determining the bolometric correction of quasars (AGNs) ? ? ! • Accretion disk models: distinguish thin or slim disks Eddington Accretion Ratio:

Modern Quasar SED Our project • SEDs from radio to X-ray utilizing best data from both ground and space telescopes • Update Elvis et al. 1994 • Better estimate of bolometric luminosity and correction • Multi-wavelength study of quasar physics

AGN SEDs: Revisit Spitzer Hubble FUSE Chandra XMM Compton gamma-ray Observatory VLA surveys Sub-mm array Optical Ground-based

Sample (heterogeneous) Total 85 quasars from 3 sub-samples: • Sub-sample 1: 22 PG quasars (a complete sample) (Laor et al. 1994, Shang et al. 2003) • Sub-sample 2: 17 AGNs from FUSE UV-bright sample (Kriss 2000, Shang et al. 2005) • Sub-sample 3: 50 radio-loud quasars (Wills et al. 1995, Netzer et al. 1995) Ø Low-redshift, z < 0. 5 (most) Ø Quasi-simultaneous UV-optical spectra to reduce uncertainty from variability

Data (UV-optical) • Quasi-simultaneous UV-optical spectra • Rest wavelength coverage 1000 – 8000 Å, (some 900 – 9000 Å) FUSE HST ground-based

Data (Infrared) • 2 MASS near-IR JHK photometry • Spitzer IRS mid-IR spectra (rest frame ~5 -35 µm) • MIPS far-IR (24, 70, 160 µm) photometry IRS spectra: • Silicates features at 10 and 18 µm (Siebenmorgen et al. 2005, Sturm et al. 2005, Hao et al. 2005, Weedman et al. 2005) • Emission lines [Ne III]15. 56 µm, [O IV]25. 89 µm, …… • Power-law between ~5 -8 µm, and beyond

Data (Radio, X-ray) Radio • Surveys from 74 MHz to 15 GHz, including 4 C, VLSS, WENSS, Texas, FIRST, NVSS, GB 6, and some GHz surveys • Higher resolution allows to separate the real cores for some objects X-ray • Chandra + XMM archive data/literature • Higher resolution and sensitivity

Spectral Energy Distribution Radio to X-ray

Spectral Energy Distribution Radio to X-ray Compared with Elvis 1994: similar in overall shape

Spectral Energy Distributions (mid-IR, optical, UV) A sub-sample of 15 objects (6 radio-loud, 9 radio-quiet) • Composite spectrum (UV + optical + mid-IR) • Normalized at 5600 Å • Clear Silicates features around 10 and 18 µm

Spectral Energy Distributions (mid-IR, optical, UV) A sub-sample of 15 objects (6 radio-loud, 9 radio-quiet) • Composite spectrum (UV + optical + mid-IR) • Normalized at 5600 Å • Clear Silicates features around 10 and 18 µm • Near-IR composite spectrum (Glikman et al. 2006) • 27 AGNs (z<0. 4) • 1 micron inflexion

Spectral Energy Distributions (mid-IR, optical, UV) A sub-sample of 15 objects (6 radio-loud, 9 radio-quiet) • Composite spectrum (UV + optical + mid-IR) • Normalized at 5600 Å • Clear Silicates features around 10 and 18 µm • Near-IR composite spectrum (Glikman et al. 2006) • 27 AGNs (z<0. 4) • 1 micron inflexion Compared to the mean SEDs of Elvis et al. 1994 (Normalized to UV-optical) • Overall similar patterns • More details with emission features

Spectral Energy Distributions (radio-loud/quiet) Normalized at 5600 Å Normalized at 8 µm Small difference between radio-loud and radio-quiet in mid-IR

Spectral Energy Distributions (diversity) • Individual mid-IR spectral are different. • Contribute differently to the bolometric luminosity (LMIR~8% to 30% of LBol, assuming LBol=9λLλ(5100Å) Normalized at 8 µm Normalized at 5600 Å

Spectral Energy Distributions => Bolometric Luminosity • In progress … • Bolometric luminosity estimate must take into account the diversity of the (mid-) infrared spectra. • Mid-IR spectra can help to improve the bolometric correction, e. g. , • Two problems: 1. Host galaxy contamination 2. Double counting

Bolometric Luminosity (2 problems) 1. Host galaxy contamination up to 50% or more in near-IR Mc. Leod & Rieke 1995 Can be corrected with high-resolution imaging of host galaxies.

Bolometric Luminosity (2 problems) 1. Double Counting • This problem can NOT be solved without assumptions. • The bolometric luminosity is an upper limit.

Conclusions • Quasar SEDs, bolometric luminosity and bolometric corrections are important. • It is hard to do. • We must do it. Thank you !

Result 2 of 3: Evidence of Intrinsic Reddening

Result 2 of 3: Evidence of Intrinsic Reddening (Is it real? ) • Correlation holds without the “outliers”.

Result 2 of 3: Evidence of Intrinsic Reddening (is it real? ) • Correlation holds without the “outliers” • Correlation is NOT caused by a correlation between spectral slope and the UV luminosity. • Show direct evidence of intrinsic dust reddening. • All quasars have intrinsic reddening (our sample is blue). • Mid-IR + UV-optical info could lead to good estimate of intrinsic reddening.

Result 3 of 3: Eigenvector one (EV 1) in Mid-IR (Boroson & Green 1992)

Result 3 of 3: Eigenvector one (EV 1) in Mid-IR r=0. 64, p=1. 0% • Equivalent width of Silicates 10µm also seems to be a parameter of EV 1. • Consistent with the picture of covering factor.

Summary • We constructed the UV-optical and mid-IR composite spectra of lowredshift broad-line (type I) quasars from a sub-sample. • Unlike borad-band SEDs, the composites show detailed mid-IR features. • Mid-IR spectra needs to be considered in estimating a better bolometric luminosity. • All quasars seem to have intrinsic dust reddening. • Mid-IR and UV-optical information may be used to estimate the intrinsic reddening. • Silicates 10µm feature is a parameter in the Eigenvector 1 relationships. • This agrees with the UV-optical results.