May 19-22, 2014 Bologna Models for AGN dust tori Michael Rowan-Robinson Imperial College London

May 19-22, 2014 Bologna Unified models for AGN: first ideas first attempt at ‘unified model’ for quasars, Seyferts, radio-galaxies: Rowan-Robinson Type 2 are obscured Type 1, link to 10  m emission ? Miley et al 1984: IRAS detection of mid-ir excess from AGN. R-R & Crawford 1989: first model of this excess as dust torus around AGN developed into unified model (Antonucci 1993, Krolik 1999) in which different AGN types are a function of the viewing angle more detailed models: Pier and Krolik 1992, Granato and Danese 1994, Rowan-Robinson 1995, Efstathiou & R-R 1995, Bianchi et al 2012 today: X-ray AGN classified as: unobscured (lg N(H) 24) -these should all be detectable as dust tori in the mid-ir -open question; could there be a major population of heavily obscured AGN, not detected at 1-10 keV in X-rays (could help explain hard-X background)

Ma y , Bol ogn a AGN dust torus geometry Efstathiou & R-R 1995

AGN dust torus seds as function of viewing angle tapered disk models Efstathiou & R-R 1995 May 19-22, 2014 Bologna

3-D radiative transfer models for clumpy AGN dust tori - Rowan-Robinson 1995, slab model for clumpy torus in QSOs - Nenkova et al 2002, 2008, 5-10 clump model for NGC1068 (slab model) - Dullemond and van Bemmel 2005, Monte Carlo - Shartmann et al 2005, 2008, Monte Carlo - Honig et al 2006, Monte Carlo R-R 95 May 19-22, 2014 Bologna

Recent studies of AGN seds Franceschini et al 2005: seds of X-ray AGN in XMM MDS Polletta et al 2006, 2007: seds of X-ray AGN in Lockman Rowan-Robinson, Valtchanov & Nandra, 2009: comparison of X-ray (CLAS-X) and mid-ir (Spitzer-SWIRE) data Hatziminoglou et al 2010: Herschel observations of QSOs Hao et al 2010: weak dust tori in XMM-Cosmos Bianchi et al 2012: AGN obscuration and the unified model Ruiz et al 2013: modelling 13 HLIRGs with IRS data – all have strong tori Lusso et al 2013: seds modeled for 32 QSOs with vel. shifts

May 19-22, 2014 Bologna Spectral energy distributions (SEDs) of a deep Lockman X-ray sample (Polletta et al 2006) BlueRed Blue Red Soft Hard Soft 26% Type 1 19% Type 2 43% Normal 8% AGN/ULIRG 4% True normal (see also Franceschini et al. 2005; Rigby et al. 2004, 2005; Alonso-Herrero et al. 2004; Alexander et al. 2001, 2002; Fadda et al. 2001)

May 19-22, 2014 Bologna X-ray detected Compton-thick AGN Selection criterion: HR, z  N H ≥10 24 cm -2 5 sources (z= ) SEDs: 2 AGN (40%) 3 normal galaxies (60%) (Polletta et al 2006) other studies of Compton- thick AGN: Daddi et al 2007 Fiore et al 2008,2009 Alexander et al 2008 Brightman et al 2012 X-rayOpticalIRX-ray Optical IR νFν Observed Wavelength Stars AGN Stars AGN

May 19-22, 2014 Bologna F(0.3-8 keV) = 2.7x erg cm -2 s -1 HR = 0.85 N H = 2x10 24 cm -2 & L x = 1.2x10 46 erg s -1 L bol = 4x10 47 erg s -1 & M BH =3x10 9 M ☉ Optical: blue with narrow lines (z=2.54) scattered light (0.6%) IR: red (r’-K = 4.2) model as QSO with A V =4 The most luminous Compton-thick quasar, z=254 Fν (μJy) z=2.54 (Keck/LRIS) Fν (μJy) dust torus model (Eftstathiou and RR 2003)

May 19-22, 2014 Bologna HerMES survey

May 19-22, 2014 Bologna combining surveys by Chandra and Spitzer Rowan-Robinson, Valtchanov, and Nandra 2009, MNRAS

May 19-22, 2014 Bologna some statistics on Spitzer-SWIRE survey over 1.5 million galaxies in SWIRE survey 10% have z > 2, 4% have z > 3 20% detected at 24  m, 1% at 70 or 160  m 35% of 24  m sources are dust torus dominated at 8  m (9% of these are QSO1, most of rest Seyfert) 5% of 24  m sources are hyperluminous (Lir > ) do find some reddened QSOs, but only 5% of QSO1 have A v > 0.5

May 19-22, 2014 Bologna SWIRE+X-rays CLASX Chandra X-ray survey in Lockman (Yang et al 2004), 0.4 sq deg, 426 extragalactic X-ray sources, 322 detected by SWIRE, 162 of which are QSOs or have dust tori a further 273 SWIRE sources are QSOs or have dust tori (total SWIRE population in CLASX area: 8563) black dot: E, Sab red dot: spiral,sb blue dot: QSO green sq: dust torus LXLX

May 19-22, 2014 Bologna 24  m v. X-ray flux Wide range of ratio of 24  m flux to X-ray flux ( mJy/ erg/cm 2 /s) (Rowan-Robinson, Valtchanov, Nandra, 2009) green triangle: broad line blue square: narrow line red square:galaxy spectr. magenta square: no spec z black circle: optical QSO crosses: upper limits filled circles- ir template type red: M82 starburst blue: AGN dust torus green: Arp 2200 starburst

May 19-22, 2014 Bologna Ltor/Lopt v. Ltor for SWIRE AGN dust tori red: M82 sb dominated at 8  m blue: AGN dust torus dominated

May 19-22, 2014 Bologna Ltor/Lopt v. Lor use volume limited sample of tori to quantify covering factor f=L(tor)/L(opt-uv). Find = -0.4,  =0.26, I.e. f = 0.4. [cf Roseboom et al (2013): = -0.41,  =0.2]

May 19-22, 2014 Bologna Hydrogen column-density v. X-ray luminosity Use hard/soft X-ray flux ratio to estimate line-of- sight absorption, N(H) Z>2.3 L Xh is corrected for absorption

May 19-22, 2014 Bologna L(tor) v. L(Xh,c) luminosity in dust torus versus X-ray luminosity lines correspond to covering factor f = 0.01, 0.1, 1, 10 (can’t have > 1! So either X-ray bolometric correction underestimated, or there has been additional absorption in X-rays) 1 0.1

May 19-22, 2014 Bologna L(tor)/L(opt) v. L(opt) lg {L(tor)/L(opt)} ~ lg f, versus lg L(opt), (L sun ) for all SWIRE AGN. we see many low-luminosity tori with SWIRE which are not detected at X-rays green triangle: broad line blue square: narrow line red square:galaxy spectr. magenta square: no spec z black circle: optical QSO crosses: upper limits filled circles- ir template type red: M82 starburst blue: AGN dust torus green: Arp 2200 starburst

May 19-22, 2014 Bologna L(opt)/LX v. soft/hard lg {L(opt)/LX} versus ratio of soft to hard X-ray flux for all SWIRE AGN. green triangle: broad line blue square: narrow line red square:galaxy spectr. magenta square: no spec z black circle: optical QSO crosses: upper limits filled circles- ir template type red: M82 starburst blue: AGN dust torus green: Arp 2200 starburst excluding QSOs flagged as optically extended or with N(H) > define L bh ~27 L Xh ~2 L opt ~ 2.5 L tor

May 19-22, 2014 Bologna L(tor)/L(opt) v. LX lg {L(tor)/Lbh} ~ lg f, versus lg LX, (L sun ) for all SWIRE AGN. objects above horizontal line are candidates for being Compton thick. green triangle: broad line blue square: narrow line red square:galaxy spectr. magenta square: no spec z black circle: optical QSO crosses: upper limits filled circles- ir template type red: M82 starburst blue: AGN dust torus green: Arp 2200 starburst

May 19-22, 2014 Bologna Summary on covering factor RR et al 2009: covering factor decreases from high to moderate X-ray luminosity, then may increase towards lower luminosities Hasinger (2008): ratio of narrow-line to broad-line AGN increases with decreasing hard X-ray luminosity Maiolino et al (2007): Type 2 fraction increases as luminosity decreases (based on ratio of 6.7  m to 5100 A) La Franca et al (2005), Hasinger (2008), Treister et al (2008): suggest fraction of obscured AGN increases with redshift

May 19-22, 2014 Bologna L(tor) v. z lg {L(tor)} versus z for all SWIRE AGN. wider dynamic range than the X-ray luminosity, more low-luminosity, low-z, dust tori blue square: narrow line red square:galaxy spectr. black circle: optical QSO filled circles- ir template type red: M82 starburst blue: AGN dust torus green: Arp 2200 starburst

May 19-22, 2014 Bologna very strong (Compton thick ?) tori Left: L(tor) > L(bh) - Compton thick candidates ? Right: lg N(H) > 24 (cm -2 ) - Compton thick

May 19-22, 2014 Bologna summary on Compton thick AGN Polletta et al (2006) identifies 7 sources with N(H) > cm -2, estimated 55 in 0.6 sq deg, only 20% of which detected in X-rays. R-R, Valtchanov, Nandra (2009) found 1 source with with N(H) > cm -2, with a further 11 candidates for which L tor >L bh. Estimated > 130 in 0.4 sq deg CLAS-X area, >20% of sample. Fiore et al (2009) estimate 44% of X-ray sources with log Lx = and 67% with log Lx = are Compton thick. Alexander et al (2008) note paucity of reported Compton thick objects at z = Discuss one new object + 6 in literature.

May 19-22, 2014 Bologna very weak tori ? L(tor) < 0.03 L(bh) - very weak or no tori ? All consistent with 1-3% covering factor. No strong evidence for < 1% covering factor (Rowan-Robinson et al 2008) see also: Jiang et al 2010 (two z~6 QSOs with no detectable hot dust), Hao et al 2010 (found substantial population of ‘hot-dust-poor’ AGN in XMM-Cosmos, ~10%) - torus destroyed during merger ?

May 19-22, 2014 Bologna starburst-AGN connection Left: L(sb) (measures star-formation rate) v. L(Xh,c) [broken line - locus for X-ray starbursts]. Right: L(tor) v L(sb) - RR(2000), hyperluminous gals. (1) Strong correlation - link between AGN and star formation (2) these are differential versions of the Magorrian diagram (M bh - M * )

May 19-22, 2014 Bologna X-ray starbursts 7-10 X-ray starbursts, L X up to erg/s Rowan-Robinson, Valtchanov, Nandra, 2009 X-ray starbursts

May 19-22, 2014 Bologna Obscuration and unified model Spectra of NGC 1365 showing occultation by BLR clouds Model of absorbing radiation for right-hand event. (Maiolino et al 2010) Bianchi et al 2012

May 19-22, 2014 Bologna Obscuration and unified model Left: Model for Circinnus galaxy (Tristram et al 2007). Right: dusty torus derived from mid-ir interferometric observations

May 19-22, 2014 Bologna Obscuration and unified model Geometry of circumnuclear matter in NGC 1068 (Raban et al 2009). Red: compact dust from mid-ir interferometry. Blue: ionization cones from OIII. Yellow: ionization cones from spectroscopy.