Obscured AGN in the COSMOS field: the multiwavelength challenge

1
Obscured AGN in the COSMOS field the
multiwavelength challenge
Marcella Brusa (MPE)
Collaborators F. Civano, A. Comastri, G.
Zamorani, F. Fiore (optical ID) XMM-COSMOS
team G. Hasinger (PI), N. Cappelluti, M. Elvis,
A. Finoguenov, R. Gilli, V. Mainieri, T. Miyaji,
C. Vignali ... P. Capak, H. McCracken, D.
Thompson (optical catalogs) M. Salvato, O.
Ilbert, H. Aussel S-COSMOS (Spitzer) J. Trump
IMACS, S.Lilly zCOSMOS (redshifts)
2
Motivation of the XMM-COSMOS project
(One of the) Main goal of the XMM-Newton Wide
field survey in the COSMOS field is study
the evolution of (obscured) Active Galactic
Nuclei over the cosmic time and the dependence of
black hole growth on galaxy morphology and
environment
(Hard) X-ray selection
Massive Redshift campaigns
High resolution optical data
Large, contiguous area

• Deep multiwavelength coverage needed to properly
  study and characterize AGN population as a whole
• ? Check for selection criteria (X-ray
  vs. optical vs. IR vs. radio)
• ? Determine bolometric output obscured
  sources and low-L objects
• ? Complete picture of AGN/galaxy
  (co)-evolution

3
COSMOS major components (in order of appearance)
HST/ACS (I-band 590 orbits
I(AB)27) 2002-2003 Subaru imaging (25
nights - b,v,r,i,z26/27) VLA (265
hours 24 µJy) GALEX deep (200 ks,
AB25) XMM-Newton (800 ks 10-15 cgs)
2004-2005 XMM-Newton (600 ks)
VLT (540 hours) Magellan (12 nights)
SPITZER-IRAC (200 hours - 1 µJy) 2006
SPITZER-MIPS (200 hours - 70 µJy)
Chandra (1.8 Ms)
Cosmos Survey 2 deg2 (PI N. Scoville)
C-COSMOS 1.8 Ms (PI M. Elvis)
AO3 800 ks of XMM 25x32 ks pointings ?
reduction/analysis completed 1400 AGN
detected 70 clusters/diffuse sources
AO4 600 ks (Total 1.4 Ms) of XMM
?reduction on-going 2000 sources expected at
the completion of the survey
MAMBO, CFHT, Bolocam and (future) others
http//www.astro.caltech.edu/cosmos/
4
XMM-COSMOS in a context
-16
CDFN-CDFS 0.1deg2 Barger et al. 2003 Szokoly et
al. 2004
E-CDFS 0.3deg2 Lehmer et al. 2005
EGS/AEGIS 0.5deg2 Nandra et al. 2006
C-COSMOS 0.9 deg2
-15
ELAIS-S1 0.5deg2 Puccetti et al. 2006
Flux 2-10 keV (cgs)
XMM-COSMOS 2 deg2
HELLAS2XMM 1.4 deg2 Cocchia et al. 2006 Champ
1.5deg2 Silverman et al. 2005
-14
SEXSI 2 deg2 Eckart et al. 2006
-13
XBOOTES 9 deg2 Murray et al. 2005, Brand et al.
2005
Area
(see Brandt Hasinger 2005 review (ARAA 43,
827)
5
COSMOS source counts

• logN-logS (normalized to
• Euclidean slope)
• 0.5-2 keV
• Confirm previous results
• with unprecedent accuracy in
• the flux range 8x10-16-5x10-12 cgs
• 5-10 keV
• In between previous determinations in the
  flux range 10-14- 5x10-12 cgs
• Able to explain most of differences in source
  counts from different surveys
• as a combination of Poissonian noise and
  clustering effects

Cappelluti et al. 2007
6
From X-ray counts to bolometric properties

• (First) Crucial step is X-ray source
  identification
• ? assign redshift, investigate extinction
  properties, derive physical quantities
• COSMOS has (or will have soon) all the data to
  make the source identification at best
• Problem XMM-Newton PSF is 6
• ? optical data too deep (Scoville et al.
  2007)
• 1 object expected by chance in a 5 error
  box
• at the 24-26 mag
• ? large spread on X/O ratios for X-ray
  sources

7
Optical identificationX/O diagram

• optically faint (Igt24) ? difficult to identify
  using optical bands only see also Alexander et
  al. 2001
• 200 sources in the XMM-COSMOS sample
• candidate high-z (zgt1) obscured QSO, zgt4
  QSO
• Efficiency of combination of IRhard X-ray
  surveys Mignoli et al. 2004, Koekemoer et al.
  2004, Brusa et al. 2005, Severgnini et al. 2005,
  Maiolino et al. 2006 and many others!

1300 sources
Brandt Hasinger 2005
8
From optical bands to IRAC/MIPS
Courtesy Salvato, Ilbert S-COSMOS

• 100 objects in XMM-COSMOS identified through
  IRAC (most of them EROs/red objects/optically
  faint)
• Very hard to get redshift from optical ?
  alternative approaches ISAAC/IRS spectroscopy
  and/or SED fitting
• Koekemoer et al. 2004, Mainieri et al.
  2005, Maiolino et al. 2006
• See Antons talk

IRAC 3.6
IRAC 4.5
IRAC 8.0
ACS

• Importance of MIPS data
• for identification of optically
• faint AGN (see Fabrizios talk)

9
Examples of XMM/IRAC coincidenceson
bright/ambiguous sources

• 100 objects in XMM-COSMOS with multiple/none
  IRAC cps
• ? more accurate X-ray positions needed to
  pick up the right cp
• ? C-COSMOS

IRAC 4.5
IRAC 3.6
IRAC 4.5
IRAC 3.6
IRAC 8.0
IRAC 8.0
ACS
ACS
Courtesy Salvato, Ilbert S-COSMOS

• Both can be counterparts
• Try to put BOTH in slits
• In spectroscopic programs

10
Redshift distribution
compilation from ongoing spectroscopic projects
IMACS/zCOSMOS SDSS literature data

• 380 secure spectroscopic identifications (zgt0)
• 25 of the full sample, 50 completeness
  in the Ilt22 sample
• BL AGNs dominate at zgt1
• ? High redshift type 2 objects missing
  (partly selection effect)
• see also results from HELLAS2XMM, Cocchia
  et al. 2006 and from the SEXSI survey, Eckart
  et al. 2006

Brusa et al. 2007 ApJS
11
Optical/IR Colors of X-ray sources (1)
BL AGN occupy the classic QSO locus in UBBV
diagrams and are mostly (80) associated with
pointlike sources
BL AGN (from spectroscopy) occupy the expected
locus also in IRAC diagrams (steep PL SED)
Lacy et al. 2004 locus
Stern et al. 2005 locus
(Updated)
Brusa et al. 2007, ApJS
See also results in Heng Hao poster
12
Optical/IR Colors of X-ray sources (2)
Obscured sources (NOT BL AGN and hard sources)
are preferentially associated with red objects
IRAC colours of NL AGN (RED) and galaxies (CYAN)
show significant contribution from host galaxy
light Optically faint (igt25 or undetected)
sources have steep PL SED ? OBSCURED AGN
Lacy et al. 2004 locus
Brusa et al. 2007 ApJS
(dichotomy confirmed by spectral analysis,
Mainieri et al. 2006)
13
Summary

• Full multiwavelength coverage needed to
• - correctly identify the X-ray counterparts
  (at least for XMM!)
• - properly study and characterize AGN
  population as a whole
• First results from XMM-COSMOS
• - logN-logS excellent agreement with XRB
  modeling
• (Cappelluti et al. 2007)
• - multicolor analysis of the optically
  bright sample suggests 80 agreement between
  spectroscopic, morphological and X-ray properties
  (as expected from unified schemes) Brusa et al.
  2007
• - candidate obscured QSOs isolated through
  K/IRAC photometry among the optically
  unidentified (faint) sources

14
XMM-COSMOS on-going projects

• Hasinger2007
• Survey description and X-ray properties
• Cappelluti, Hasinger, Brusa et al. 2007
• Source counts, LogN-LogS
• Brusa, Zamorani, Comastri, Hasinger et al. 2007
• AGN identification, classification multiwave
  properties
• Mainieri2007
• X-ray spectral analysis of point-like sources
  
• Miyaji2007
• Angular Correlation Function
• Finoguenov2007
• Study of groups and clusters in XMM-COSMOS
• ? to appear in an ApJS special issue (2007)
• http//www.mpe.mpg.de/XMMCosmos/PAPERS/
• ..and many others projects on-going!

15
FAQ on XMM-COSMOS

• 1) Where do we find all the things you showed?
• Hasinger2007
• Survey description and X-ray properties
• Cappelluti, Hasinger, Brusa et al. 2007
• Source counts, LogN-LogS
• Brusa, Zamorani, Comastri, Hasinger et al. 2007
• AGN identification, classification multiwave
  properties
• Mainieri2007
• X-ray spectral analysis of point-like sources
  
• Miyaji2007
• Angular Correlation Function
• Finoguenov2007
• Study of groups and clusters in XMM-COSMOS
• ? to appear in an ApJS special issue (2007)
• http//www.mpe.mpg.de/XMMCosmos/PAPERS/
• ..and many others projects on-going!
• including X-EROs (F. Civano), 3-D correlation
  function (R. Gilli), EXOs (A. Koekemoer), photoz
  (M. Salvato), Type 1 QSO host (K. Jahnke) etc.

16
First look at ACS morphologies
Pointlike
Images 20 x 20 in size
z1.236
z2.454
z1.506
z2.008
Extended
z0.044
z0.125
z0.349
z0.365
z0.623
z0.969
z0.863
z0.552
17
COSMOS major components HST/ACS (i-band 590
orbits I(AB)27) Subaru imaging (25 nights,
b,v,r,i,z,) VLT (540 hours) Magellan (12
nights) XMM-Newton (1.4 Ms) VLA (265 hours)
GALEX deep (200 ks, AB25) SPITZER (200 hours)
all underway !
MAMBO, CFHT, TNG and others
http//www.astro.caltech.edu/cosmos/
18
Optical identificationsin XMM-COSMOS

• 1390 X-ray sources (from 23 fields sample)
• Identification in the optical (I) band using a
  statistical method to match the X-ray error box
  to the most likely optical cp (likelihood ratio
  technique, Sutherland Saunders 1992)
• Method tested and described in Brusa et al.
  ApJS paper
• Refined with the (old) K-band catalog (but
  too shallow..)
• - identified sample (optically/secure
  identified, 80)
• ? quite solid, large sample ?
  statistical properties (colors/redshifts/mo
  rphologies/selection effects)
• - ambiguous/unidentified sample (problems with
  ID, 20)
• ? most interesting sources ? candidate
  high-z, obscured QSO

19
Examples of XMM/IRAC coincidences
ACS
IRAC 3.6 micron
(Courtesy M. Salvato, D. Sanders)

• 120 objects in XMM-COSMOS identified through K
  and/or IRAC (most of them EROs)
• Very hard to get redshift from optical ?
  alternative approaches ISAAC/IRS spectroscopy
  and/or SED fitting
• Maiolino et al. 2006, Koekemoer et al.
  2004
• Cfr. Also sample of highly obscured
  objects
• in Martinez Sansigre et al. 2005

20
First morphological test on identifications
ACS morphological information
About 50 of the IDs have stellar (or almost
stellar FWHMlt3 pixels) profile on ACS data
blue points pointlike red points extended
Very soft (HR-1) sources are mostly
point-like (dashed histogram) Very hard (HR1)
sources are preferentially associated with
extended objects (solid histogram)
Brusa et al. 2006
21
U-B vs. B-V diagram
Colors of X-ray sources (1)
The majority (80) of stellar-like objects from
ACS occupy the locus of quasars in the U-B vs.
B-V diagram Conversely, gt70 of UBV selected
objects recovered in the X-rays (90-100 expected
with the completion of XMM-COSMOS) Additional
50 (not shown) classified as extended ? missed
in color selected diagrams
Brusa et al. 2006
22
R-K vs. K
Colors of X-ray sources (2)
Extended sources are significantively redder
than point-like and associated with NOT BL AGN
Very soft (HR-1) sources are mostly associated
with blue sources (dashed histogram) Hard
sources (HRgt-0.3) are preferentially associated
with red objects (solid histogram)
(AB mag)
See Francescas talk!
(dichotomy confirmed by spectral
analysis Mainieri et al. 2006)
Brusa et al. 2006