The Chandra survey of the COSMOS field Fabrizio Fiore & the C-COSMOS team Particular thanks to T. Aldcroft, M. Brusa, N. Cappelluti, F. Civano, A. Comastri, M. Elvis, S. Puccetti, C. Vignali, G. Zamorani M. Salvato & S-COSMOS team & many others

Table of content  Presentation of the survey  C-COSMOS in a context  Selected scientific results  Close pairs  High-z QSOs  Fraction of obscured AGN  Summary (what you should bring home after all..)  Multiwavelength coverage is mandatory  X-ray is the leading band for all AGN studies (provided that X-ray coverage is deep enough)  Presentation of the survey  C-COSMOS in a context  Selected scientific results  Close pairs  High-z QSOs  Fraction of obscured AGN  Summary (what you should bring home after all..)  Multiwavelength coverage is mandatory  X-ray is the leading band for all AGN studies (provided that X-ray coverage is deep enough)

Co-evolution of galaxies and SMBH Two seminal results: 1.The discovery of SMBH in the most local bulges; tight correlation between M BH and bulge properties. 2.The BH mass density obtained integrating the AGN L.-F. and the CXB ~ that obtained from local bulges  most BH mass accreted during luminous AGN phases! Most bulges passed a phase of activity: 1)Complete SMBH census, 2) full understanding of AGN feedback are key ingredients to understand galaxy evolution

The C-COSMOS survey: which science  Black hole growth and census  XMM has ~20% of ambiguous identifications. Chandra survey secures the discovery and identifications of rare objects (elusive AGN, high-z AGN).  The combination of Chandra data and Spitzer’s 24  m and 3-8  m data allows us to unveil highly obscured accretion, thus providing a complete census of accreting SMBH  The influence of the environment on galaxy activity  excesses of X-ray point sources (AGN) within a few Mpc of clusters at 0.2<z<1.  spikes in the redshift distribution of the X-ray sources  The AGN and galaxy ACF and CCF down to a few arcsec: how the AGNs trace the cosmic web.  AGN pairs with separation<10-20”: galaxy activity vs. galaxy interaction  Black hole growth and census  XMM has ~20% of ambiguous identifications. Chandra survey secures the discovery and identifications of rare objects (elusive AGN, high-z AGN).  The combination of Chandra data and Spitzer’s 24  m and 3-8  m data allows us to unveil highly obscured accretion, thus providing a complete census of accreting SMBH  The influence of the environment on galaxy activity  excesses of X-ray point sources (AGN) within a few Mpc of clusters at 0.2<z<1.  spikes in the redshift distribution of the X-ray sources  The AGN and galaxy ACF and CCF down to a few arcsec: how the AGNs trace the cosmic web.  AGN pairs with separation<10-20”: galaxy activity vs. galaxy interaction

The C-COSMOS survey: how  The Chandra high resolution permits to resolve sources 2” apart over 0.9 sq. deg., corresponding to 8-16 kpc separations for z = , and locates point sources to < 4 kpc at any z. Thus close mergers can be resolved, and AGNs can be distinguished from ULXs and off-nuclear starbust.  Thanks to the good PFS, ACIS-I is not background limited, then C-COSMOS reaches ~3 times deeper than XMM- COSMOS in both hard and soft bands and cross the threshold where starburst galaxies become common in X- rays.  The low ACIS background enables stacking analysis, in which counts at the positions of known classes of objects are co-added to increase the effective exposure time  The Chandra high resolution permits to resolve sources 2” apart over 0.9 sq. deg., corresponding to 8-16 kpc separations for z = , and locates point sources to < 4 kpc at any z. Thus close mergers can be resolved, and AGNs can be distinguished from ULXs and off-nuclear starbust.  Thanks to the good PFS, ACIS-I is not background limited, then C-COSMOS reaches ~3 times deeper than XMM- COSMOS in both hard and soft bands and cross the threshold where starburst galaxies become common in X- rays.  The low ACIS background enables stacking analysis, in which counts at the positions of known classes of objects are co-added to increase the effective exposure time

C-COSMOS in a context HST ACS imaging HST ACS imaging with resolution 0.05” and sensitivity 27.2 mag (10  ) provides morphologies of over 2 milions galaxies at < 100 pc resolution! IR/Optical/UV large surveys to improve photometric redshift Spitzer: IRAC-deep MIPS-Shallow MIPS-Deep zcosmos Optical spectroscopy surveys: zcosmos :540 hours on the ESO VLT using VIMOS. Magellan COSMOS VLA-Cosmos Large Project plus submm XMM-Newton XMM-Newton : 1.4 Msec. Chandra! Chandra! Cycle 8 proposal 1.8 Msec 200ksec 0.9sq.deg f lim ~2x cgs (0.5-2 keV)

40 arcmin 52 arcmin z = 0.73 struct ure z-COSMOS faint Color: XMM first year Full COSMOS field C-COSMOS: numbers  1.8 Ms total exposure time  36 ACIS-I pointings  200 ksec average exposure 0.5deg 2  100 ksec average exposure 0.4deg 2  F lim ~2x cgs (0.5-2 keV)  1759 X-ray sources (probability threshold 2x10 -5 )  1.8 Ms total exposure time  36 ACIS-I pointings  200 ksec average exposure 0.5deg 2  100 ksec average exposure 0.4deg 2  F lim ~2x cgs (0.5-2 keV)  1759 X-ray sources (probability threshold 2x10 -5 ) Elvis et al. 2008

The C-COSMOS multiwavelenth catalog  Identification in the 3.6micron K, and I bands using a statistical method to match the X-ray error box to the most likely cp (“likelihood ratio technique”)  “identification” in 3 bands sample: 94% !!  IR “identified” sample 5%  most interesting sources  high-z QSOs, obscured QSOs  ambiguous/unidentified sample 1%  870 sources in common with XMM 895 NEW sources!!  450 spectroscopic redshift already in hand(SDSS,VIMOS,IMACS)  Photometric redshift already available for 60% of the sample  Identification in the 3.6micron K, and I bands using a statistical method to match the X-ray error box to the most likely cp (“likelihood ratio technique”)  “identification” in 3 bands sample: 94% !!  IR “identified” sample 5%  most interesting sources  high-z QSOs, obscured QSOs  ambiguous/unidentified sample 1%  870 sources in common with XMM 895 NEW sources!!  450 spectroscopic redshift already in hand(SDSS,VIMOS,IMACS)  Photometric redshift already available for 60% of the sample Obscured AGN unobscured AGN SFgalaxies XBONGs Star Extreme AGN 5% XMM-COSMOS limit on 1deg 2 Civano et al 2008

Close pairs Thanks to the good Chandra PSF it is possible to study close pairs to search for X-rays from galaxy interactions. Wavelet detection algorithm (PWDETECT, Damiani et al.) optimized to resolve nearby sources (Puccetti et al. 2008). A total of 106 sources closer than 12” are present in the X-ray catalog. > than expected from simulation. Next step is to obtain the spectroscopic identification to verify the fraction of physical pairs (Vignali et al. 2008)

Chandra/XMM comparison 50% of the chandra pairs have associated only one XMM source. In several cases the brightness of the sources of the pair is similar. BLUE circles= keV chandra detections. Green =XMM contours

High redshift AGN C-COSMOS XMM-COSMOS Elvis et al Brusa et al Civano et al C-COSMOS XMM-COSMOS Elvis et al Brusa et al Civano et al XMM-COSMOS: QSO z>3 ~30 deg 2 QSO z>4 ~3 deg 2 Chandra ~3 times deeper than XMM QSO z>3 deg QSO z>4 deg 2

Obscured AGN ChandraU ACS K 3.6  m 4.5  m Type 1 AGN Non type 1 AGN MIR/O>1000 High X/O, high MIR/O

Evidences for missing SMBH While the CXB energy density provides a statistical estimate of SMBH growth, the lack, so far, of focusing instrument above 10 keV (where the CXB energy density peaks), frustrates our effort to obtain a comprehensive picture of the SMBH evolutionary properties. Gilli et al Marconi Menci, Fiore et al. 2004, 2006,

AGN density > La Franca, Fiore et al Menci, Fiore et al Paucity of Seyfert like z>1 is real? Or, is it, at least partly, a selection effect? Are we missing in Chandra and XMM surveys highly obscured (N H  cm -2 ) AGN? Which are common in the local Universe… Paucity of Seyfert like z>1 is real? Or, is it, at least partly, a selection effect? Are we missing in Chandra and XMM surveys highly obscured (N H  cm -2 ) AGN? Which are common in the local Universe…

Why multiwavelength surveys  IR surveys:  AGNs highly obscured at optical and X- ray wavelengths shine in the MIR thanks to the reprocessing of the nuclear radiation by dust  IR surveys:  AGNs highly obscured at optical and X- ray wavelengths shine in the MIR thanks to the reprocessing of the nuclear radiation by dust Dusty torus Central engine

7m7m 7.7  m 800pc 100pc Laurent et al. 01 IR surveys  Difficult to isolate AGN from star-forming galaxies (Lacy 2004, Barnby 2005, Stern 2005, Polletta 2006 and many others)

 Use both X-ray and MIR surveys:  Select unobscured and moderately obscured AGN in X-rays  Add highly obscured AGNs selected in the MIR  Simple approach: Differences are emphasized in a wide-band SED analysis  Use both X-ray and MIR surveys:  Select unobscured and moderately obscured AGN in X-rays  Add highly obscured AGNs selected in the MIR  Simple approach: Differences are emphasized in a wide-band SED analysis Why multiwavelength surveys

MIR selection of CT AGN ELAIS-S1 obs. AGN ELAIS-S1 24mm galaxies HELLAS2XMM CDFS obs. AGN Fiore et al Open symbols = unobscured AGN Filled symbols = optically obscured AGN * = photo-z Unobscured obscured X/0 MIR/O

MIR selection of CT AGN COSMOS X-ray COSMOS 24um galaxies R-K Fiore et al. 2008a Fiore et al. 2008b Open symbols = unobscured AGN Filled symbols = optically obscured AGN * = photo-z CDFS X-ray HELLAS2XMM GOODS 24um galaxies

Template highly obscured QSOs  IRAS  High L(IR)/Lx ratio  No PAH emission features in IRS spectrum  IR SED dominated by the AGN Abel2690#75 (Pozzi et al 2007) IRAS 09 SDSS spectrum

COSMOS MIR AGN Fiore et al. 2008b not directly Stack of Chandra images of MIR sources not directly detected in X-rays

AGN fraction Chandra survey of the Bootes field (5ks effective exposure) Brand et al assume that AGN populate the peak at F24um/F8um~0 only. They miss a large population of obscured AGN, not detected at the bright limits of their survey.

AGN fraction Caputi et al La Franca et al keV

CT AGN volume density A B C GCH 2007 logN H >24 z= : density IR-CT AGN ~ 45% density X-ray selected AGN, ~90% of unobscured or moderately obscured AGN z= : density IR-CT AGN ~ 100% density X-ray selected AGN, ~200% of unobscured or moderately obscured AGN The correlation between the fraction of obscured AGN and their luminosity holds including CT AGN, and it is in place by z~2 No AGN feedback AGN feedback Gilli et al model La Franca et al. 2005

AGN obscuration, AGN feedback and star-formation  CT absorbers can be naturally included in the Menci et al. feedback scenario as an extension toward smaller distances to the nucleus where gas density can be high.  If this is the case and if the fundamental correlation between the fraction of obscured AGN and L is due to different timescales over which nuclear feedback is at work  Evolutionary star-formation sequence:  CT moderately obscured unobscured  Strong moderate small  CT absorbers can be naturally included in the Menci et al. feedback scenario as an extension toward smaller distances to the nucleus where gas density can be high.  If this is the case and if the fundamental correlation between the fraction of obscured AGN and L is due to different timescales over which nuclear feedback is at work  Evolutionary star-formation sequence:  CT moderately obscured unobscured  Strong moderate small

AGN obscuration, AGN feedback and star-formation  COSMOS  Log(L5.8/L1.4GHz)=4.74 (0.12) 38 CT QSOs z=  Log(L5.8/L1.4GHz)=5.07 (0.13) 25 QSOs z=  X-ray selected, type-2 QSO have higher submm detection rate than unobscured QSO  COSMOS  Log(L5.8/L1.4GHz)=4.74 (0.12) 38 CT QSOs z=  Log(L5.8/L1.4GHz)=5.07 (0.13) 25 QSOs z=  X-ray selected, type-2 QSO have higher submm detection rate than unobscured QSO Page et al Stevens et al unobscured obscured

Density of Obscured AGNs Dashed lines = Menci model, no AGN feeback Solid lines = Menci model, AGN feedback 2-10 keV data = La Franca, FF et al Spectroscopic confirmation: very difficult for the CDFS-GOODS sources (R~27, F(24um)~100uJy Possible for the COSMOS sources!! F24um~1mJy ==> Spitzer IRS AO5 program (Pri. C, Salvato et al.) ?

Summary  Chandra sensitive survey of the COSMOS field: 1758 sources, ~half new, I.e. not detected by XMM  ~100 sources with optical counterpart fainter than I=26.5: ==> highly obscured QSOs, high-z QSOs  Large sample of bright pairs: ==> galaxy interaction vs. galaxy activity  Combined use of Chandra and Spitzer over a large field: ==> discovery of CT type 2 QSOs at z=1-2 ==> fraction of X-ray detected and X-ray emitting AGN in 24um samples is large (~50%) ==> fraction of X-ray detected and X-ray emitting AGN in 24um samples is large (~50%)  All this will allow a precise determination of the evolution of the accretion in the Universe, a precise census of accreting SMBH  While multiwavelength coverage is mandatory, X-ray is the leading band for AGN studies (provided that X-ray coverage is deep enough)  Chandra sensitive survey of the COSMOS field: 1758 sources, ~half new, I.e. not detected by XMM  ~100 sources with optical counterpart fainter than I=26.5: ==> highly obscured QSOs, high-z QSOs  Large sample of bright pairs: ==> galaxy interaction vs. galaxy activity  Combined use of Chandra and Spitzer over a large field: ==> discovery of CT type 2 QSOs at z=1-2 ==> fraction of X-ray detected and X-ray emitting AGN in 24um samples is large (~50%) ==> fraction of X-ray detected and X-ray emitting AGN in 24um samples is large (~50%)  All this will allow a precise determination of the evolution of the accretion in the Universe, a precise census of accreting SMBH  While multiwavelength coverage is mandatory, X-ray is the leading band for AGN studies (provided that X-ray coverage is deep enough)