David Elbaz - IR view of AGNs 1

1
The infrared view of obscured AGNs from Spitzer
to Herschel

• What can we learn about
• The obscured sources responsible for the peak
  emission of the CXB at 30 keV (in flux/area,
  z1)
• The growth of supermassive black holes through
  the Hubble time (in intrinsic luminosity, z2)
• David Elbaz (CEA Saclay)
• based on 2 studies at
• z1 with D.Alexander, R.R. Chary
• z2 Daddi et al. 07 paper, with Dave Alexander
  (Durham), Mark Dickinson (NOAO), Roberto Gilli
  (U. di Bologna), A.Renzini (U. di Padova)

2
Unified scenario Compton Thick AGNs edge-on
dusty torusCan IR surveys reveal the missing AGN
population unidentified in X-ray ?

• nearest radio galaxy, 3.5 Mpc
• Unified scenario dust obscuration
• depends on the orientation only.

3
IR signature of the presence of an AGN

• LIR Star formation AGN
• LIRlt2.5x1012 L? 82-94 6-18
• LIRgt2.5x1012 L? 45-55 45-55

4
Infrared properties of local star forming galaxies

• Correlations between mid and far IR
• used to derive Lbol, hence SFR.
• Far IR view of the Universe strongly affected by
  confusion due to large PSF -gt use of mid IR

5
20-30 of the CXB due to Compton Thick AGNsmost
of which remain unidentified What can we learn
from the IR ?
CT AGNs present a mid IR excess. can IR surveys
identify the missing AGNs ?
6
24 ?m
GOODS Great Observatories Origins Deep Survey
(PI Dickinson) Spitzer-IRAC MIPS 10 x 15 50
complete _at_ 30 mJy 70 complete _at_ 70 mJy
15'
10' 10 comoving-Mpc at z1
7
BzK selection massive galaxies at 1.4 lt z 2.5
(Daddi et al. 04)

• Sample of 1200 BzK selected galaxies in GOODS-S
  (down to K22 Vega)

8
Radio UV corrected for extinction -gt same LIR
consistent with star formation origin
9
Luminous objects at 8 ?m, z2 present a mid
infrared excess inconsistent with a star
formation origin
10
Luminous objects at 8 ?m, z2 present a mid
infrared excess inconsistent with a star
formation origin
11
X-ray stacking analysis (by Dave Alexander) in
GOODS-S to K22 (Vega)
Normal galaxies
MIR excess gals
mid-IR excess sources are much harder than normal
ones N.B. all direct X-ray detections and power
law sources are excluded
12
Sub-band stacking with Chandra reveals a shape
that strongly supports NH1024 cm-2 or perhaps
more, i.e. Compton Thick At 0.5-2 keV 2
populations identical consistent with star
formation origin with similar SFRs
13
Sub-band stacking with Chandra gt Compton Thick
AGNs Subtracting power-law from SF --gt
remarkably similar to AGN template with NH
3x1024 cm-2 dashed line AGN SED from Gilli et
al 07 including 6.7 keV Fe line and for the z
distribution of the sample.
14
GOODS ultra-deep mid IR survey can detect NGC
1068-like CT AGNs to z2Chandra only to z0.5
even with 2 mega-seconds !
GOODS
zmaxIR 2
zmaxX-ray 0.5
15
Contribution to the Cosmic X-ray Background

• About 20-30 of the 10-30 keV background is due
  to unidentified CT AGNs (Gilli et al. 07)
• CT AGNs with mid IR excess at z2 3200
  objects/deg2
• produce 10-25 of the missing hard X-ray
  background
• Most of the missing background comes from similar
  sources but at z1
• (negligible contribution to the 2-8 keV band)

16
Spitzer 24 ?m sources at z1 anomalously faint at
70 ?m
NGC 1068 (Le Floc'h et al. 01 70 MIR from
AGN) X-ray Stacking hard src 20 Compton Thick
AGNs ! (X-ray unselected as AGNs)
-gt 720/sq.deg. lt-gt 2800/sq.deg. required to
produce the X-ray bkg (Worsley et al. 05) ? radio
excess X-ray undetected AGNs ( Donley, Rieke et
al. 06)
17
The census of Compton Thick AGNs

• Contribution of mid IR excess sources to the
  volume density of C.T. AGNs
• at z2 consistent with 100 of CT AGNs with
  LXgt1042 erg s-1
• at z1 consistent with 25 of AGNs producing
  the CXB
• and 100 of CT AGNs with LXgt3x1042 erg s-1

18
Discussion conclusions

• Spitzer and soon Herschel are efficient at
  detecting Compton Thick AGNs missed in X-ray
  surveys.
• These "mid IR excess Compton Thick AGNs",
  confirmed from X-ray stacking, potentially
  represent a large fraction of the BH growth, but
  resolve at present 25 of the CXB assuming model
  extrapolation to 30keV.
• Official Herschel sensitivity limits will only
  marginally improve this result but actual
  expectations may do much better.

19
Discussion conclusions

• Spitzer and soon Herschel are efficient at
  detecting Compton Thick AGNs missed in X-ray
  surveys.
• These "mid IR excess Compton Thick AGNs",
  confirmed from X-ray stacking, potentially
  represent a large fraction of the BH growth, but
  resolve at present 25 of the CXB assuming model
  extrapolation to 30keV.
• Official Herschel sensitivity limits will only
  marginally improve this result but actual
  expectations may do much better.
• This population is at the level of
• 2x10-2 ?Crab at z1 and 5x10-3 ?Crab at z2
• 25 100 times fainter than expected SIMBOL X
  depth
• (0.5 ?Crab3x10-14 erg s-1 cm-2 at 10-40 keV)
• or 2.5 10 times fainter if 3x10-15 erg s-1
  cm-2 is reached

20
No optical spectral lines typical of hidden AGNs
(GMASS deep spectra C. Halliday et al)
CIII1909 (EW1-2A) similar to star forming
galaxies HeII1640 very faint