Nessun titolo diapositiva

1
AGN IN THE SPITZER ERA UNVEILING OBSCURED
ACCRETION
Carlotta Gruppioni (INAF-OABO)
Bologna, 8 Giugno 2007
2
SUMMARY

• Why infrared for detecting AGN?
• - Unobscured AGN
• - Obscured AGN
• AGN emission in the infrared
• Old results from IRAS ISO
• What is Spitzer
• - IR selection techniques for AGN
• - Recent Results from Spitzer
• How many AGN in complete IR samples?

3
WHY INFRARED?

• Infrared explores the hidden Uiverse (obscured
  by dust)
• A large fraction of soft X-ray, UV and optical
  emission from AGN is absorbed by dust and
  re-radiated in the infrared (IR)
• ?
• Sensitive measurements in the IR provide an
  opportunity to look for obscured AGN not
  identified in X-ray and optical surveys

4
Selection of unobscured AGN

• Most efficient way Hard X-ray surveys

Examples high X/O sources (moderate obscured
AGN at z1-2 hosted in massive ellipticals)
BUT hard X-ray surveys still miss the
highly obscured sources (i.e. Compton thick
Nhgt1024 cm-2) dont sample the XRB peak
Fiore et al. 2003, AA Mignoli et al. 2004,
AA Mainieri et al. 2005, AA Maiolino et al.
2006, AA
Missing population (numerous) moderately
luminous, NHgt23, z0.5-2 AGN (Worsley et al.)
5
Why should we care about obscured AGN?

• Obscured AGN are needed
• - to reproduce the X-ray background peak
  (Setti Woltjer 1989, Comastri et al. 1995 etc.
  )
• - Unified Models dusty torus around AGN
  responsible for absorption of X-ray to NIR
  nuclear radiation
• - models predict that AGN activity in the past
  take place in dusty environments/systems
• Need to separate AGN from stellar activity to
• - Have a complete picture of galaxy(-AGN)
  formation and (co)evolution
• ? only through a precise census of AGN
  activity in the Universe it is possible to obtain
  strong constraints to models for the formation
  and evolution of structure

6
AGN emission in the IR
unabsorbed

• ? The obscured AGN IR Spectral Energy
  Distribution (SED) is dominated by thermal
  emission from dust (Neugebauer 1979), heated by
  the primary optical/UV continuum dominating the
  unobscured AGN SEDs.
• ? The broadness of the IR SED, which could not be
  reproduced by means of single-temperature
  blackbody, was explained in terms of multiple
  temperature component dust.
• ? According to the AGN Unified Model, the dust is
  distributed as a (clumpy?) annular ring (torus)
  around the central BH.
• The different orientations of the dusty torus
  with respect to the line of sight are responsible
  for the differences observed in the
  X-rays/optical spectra/SEDs/etc of type 1 and
  type 2 AGN.

absobed
7
Type 2 AGN

• Optical emission from Type 2 AGN can be diluted
  by the host galaxy
• difficult to classify as AGN from optical line
  diagnostics.
• Crucial range near-/mid-infrared (NIR/MIR)
  where galaxy SEDs have a dip (2-5 ?m) and the
  hot dust heated by the AGN start contributing,
  filling up the dip.

flat
dusty torus emission
dip
8
Complementary approach IR colour selection

• AGN (unobs and obs) are expected to have
• warm power-law SEDs at gt1 µm
• (? from elliptical/starburst)

Blue (unobs)
Red
AGN
Red (obs)
AGN (both type 1 and 2) can be isolated in
NIR/MIR diagrams
?S?
Blue
Red
SEVERAL IR colour-selection criteria proposed so
far (i.e. Lacy et al. 2005 Stern et al. 2005
Barmby et al. 2006, etc.) ? PROBLEMS Completeness
(are all AGN selected?) Reliability (are only
AGN selected? How much galaxy contamination?)
Elliptical
Flat/Blue
Red
Starburst
Optical NIR IRAC 3.6
4.5 5.8 8.0
NEED Complete Multiwavelength Characterization
9

AGN in the IR from IRAS to ISO and SPITZER

• IRAS has explored the local Universe (z lt 0.2)
  in the IR (10 200 µm) for the first time,
  finding that
• The majority of the most luminous galaxies in
  the Universe emit the bulk of their energy in the
  far-infrared. They often contain both extreme
  starbursts and AGN.
• Even normal spiral galaxies radiate at least
  30 of their energy in the IR (Sanders Mirabel,
  1996, ARAA, 34, 749)
• About 13 of local 12-µm IRAS sources have a
  Seyfert 1 /Seyfert 2 nucleus (Rush, Malkan
  Spinoglio, 1993, ApJS, 89,1)

10

• MIR/FIR emission 4x stronger in the quasars than
  in the radio galaxies. Do radio galaxies contain
  potent sources of radiation that are hidden at
  shorter wavelengths? (Heckman, Chambers
  Postman, 1992, ApJ, 391, 39)
• ?
• First torus models to reconcile
• IR observations for type 1 /
• type 2 AGN with Unified Models
• (i.e. Pier Krolik, 1992, ApJ, 401, 99
• Granato Danese, 1994, MNRAS, 268,
• 235 etc.)

11
Solving the Radiative Transfer Equation
The equilibrium temperature (and hence thermal
emission) of the grains of each species in each
sample element is found by solving the thermal
equilibrium equation Qabs, Qem grain
absorption/emission efficiency B(?, Tik)
blackbody emission Tik temperature of a given
grain within the ikth sample element Jik
incoming specific intensity on the volume
element For the first iteration ?ik (?)
optical depth between the central source and the
ikth element rik distance At the second
iteration, the new incoming radiant flux is

12
Infrared Space Observatory
ISO (0.65-m Telescope) (gt 1000 times better
sensitivity than IRAS) has explored for the
first time the Universe at z gt 0.5 in the MIR
(15 µm ISOCAM) ? First deep MIR
surveys for distant galaxies and AGN
13
Main ISOCAM results on AGN

• ISO SWS Spectroscopy powerful tool to
  distinguish star-formation (PAHs) from AGN
  activity (continuum)
• (i.e. Laurent et al., 2000, 359, 887)
• AGN contribution to MIR surveys derived from
• ? Correlation analysis of deep X-ray and MIR
  observations (Lockman Hole HDF-N) AGNs are
  15-20 of the MIR population down to 0.05 mJy
  (Fadda et al., 2002, AA 383, 838)

14

• ? Optical Spectroscopy
• (ELAIS-S1) AGNs (type 1 2)
• account for 25 of the MIR
• sources down to 0.5 mJy.
• (La Franca et al., 2004, AJ, 127, 3075)
• AGN 15 µm Luminosity Function
• type-1 pure luminosity evolution
  L(z)L(0)(1z)2.9
• type-2 L(z)L(0)(1z)1.8-2.2

15
From ISOCAM to Spitzer ...
Spitzer Telescope (0.85-m) is now providing new
insight into the IR population of galaxies and AGN
In particular with the MIPS 24-?m band, detecting
the high-z (z1.5-3.0) analogs of the 15-?m
sources
16
About SPITZER
Telescope 0.85-m launch August 2003 Mission
2.5-5 years Wavelength 3 - 180 ?m Capability
Imaging/Photometry 3-180
Spectroscopy 5-40 ?m
Spectrophotometry 50-100
17
SPITZER Measurements - Imaging
18
SPITZER Measurements - Spectroscopy
19
Main Results on AGN from Spitzer

• IRAC/MIPS colour-colour diagrams are able to
  isolate unobscured/obscured AGN (i.e. Lacy et al.
  2004 Stern et al. 2005 Barmby et al. 2006
  Martinez-Sansigre et al. 2005 Fiore et al. 2007)
• IRS spectroscopy of MIR objects optically faint
  (or invisible) able to identify highly obscured
  AGN at high z (i.e. Houck et al. 2004, 2005
  Highdon et al. 2004 Weedman et al. 2006)
• Broad-band SED analysis able to identify AGN
  signature in the MIR (i.e. Alonso-Herrero et al.
  2006 Rodighiero et al. 2007 Pozzi et al. 2007
  Gruppioni et al. 2007)

20
Lacy et al., 2004, ApJS 154, 166MIR properties
of SDSS quasars (FLS survey)

• SDSS quasars (crosses) populate a
• well-defined region in the IRAC
• colour-colour space independent
• and not consecutives - and have
• S(8.0)1 mJy
• Candidate AGN are IRAC sources in
• the same locus and with S(8.0)gt1 mJy
• 16 SDSS QSO
• 16 Candidate obscured AGN
• 11 not identified/classified
• Ratio obs/unobs11
• Reverse argument
• In the same region there are 2000
• additional sources with S(8.0)lt1 mJy!!

2000 sources
21
Stern et al. 2005, ApJ, 631, 163MIR properties
of AGN in theBootes/AGES survey

• Spectroscopically confirmed AGN occupy a region
  in the IRAC color-color space - independent
• and consecutives colors
• Contamination from other objects is very low
• Selection effect Bootes/AGES survey probe only
  zlt0.6 galaxies
• QUITE SHALLOW
• (limited mainly to Rlt19-21)
• dont sample zgt0.6 galaxies!

22
Martinez-Sansigre et al. 2005, Nature 436,
666Selection type-2 AGN at z2 with Spitzer

• S(24 micron) gt 300 muJy
• sample QSO with Lgt0.2L at
• z2
• S(3.6 micron) lt 45 muJy
• remove naked type 1 AGN
• and low-z type-2
• 350 muJy lt S(1.4GHz) lt 2 mJy
• ensure candidates being
• radio-quiet QSO rather than
• SB and filter out radio-loud
• objects
• 21 candidates
• 10 spectroscopically confirmed
• at z1.4-4.2, with no BL!
• The remaining are blank spectra
• (ellipticals??)

No evidence for contamination in this sample (but
they are looking for specific objects..)
23
Fiore et al. 2007, ApJ, submittedUnveiling
Obscured Accretion in the Chandra Deep Field South
Obscured AGN z0-4 (Pozzi et al. 2007)

• High 24-µm/optical ratio (gt 1000) red colours
  (R-K gt 4.5)
• highly obscured AGN
• at high z
• Analysis of the X-ray properties of these
  extreme sources ? most are indeed likely Compton
  thick AGN

Arp 220 z0-4
Elliptical z0-4
24
MIR AGN
F24/FR gt 1000 R-K gt 4.5
F24/FR gt 1000 R-K lt 4.5
F24/FR lt 100 R-K gt 4.5
25
F24µm/FR gt1000 R-Kgt4.5
F(0.3-1.5keV)10-17 cgs
F(1.5-4keV)10-17 cgs
26
F24µm/FR lt100 R-Kgt4.5
F(0.3-1.5 keV)2?10-17 cgs
F(1.5-4 keV) lt 5?10-18 cgs
27
Highdon et al. 2004, ApJS, 154, 174 Houck et
al. 2005, ApJ, 622, L108Spitzer-IRS Spectroscopy
of optically obscured MIR sources

• Redshifts derived for 24-µm sources that are
  optically very faint (R gt 24.5 mag) from strong
  silicate absorption feature (_at_9.7 µm) ? obscured
  nuclei at 1.7 lt z lt 2.8

?
28
Need for Multiwavelength Surveys ( possibly MIR
Spectroscopy !)

• 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

29
How to quantify AGN (both obscured and
unobscured) versus galaxy fraction in complete /
statistical samples?

• Study broad-band SEDs and multi-wavelength
  properties of complete samples (i.e. MIR / X-ray
  / morphologically selected) with available
  optical spectroscopy
• confirm with MIR spectroscopy

30
Alonso-Herrero et al. 2006, ApJ 640, 167Infrared
power-law galaxies in the CDFS AGN and ULIRGs

• 24 µm sources with power-law emission in the IRAC
  3.6-8 µm bands
• Those detected in X-rays (50) have luminosities
  typical of AGN
• Those not detected in X-rays have global X-ray to
  MIR SED properties that make them good IR
  bright/X-ray absorbed candidate AGN

31
Rodighiero, Gruppioni, Civano et al. 2007,
MNRAS,376, 416 Hidden Activity in High-z
Spheroidal Galaxies from MIR and X-ray
Observations in the GOODS-N Field
168 morphologically classified (Bundy et al.
2005) spheroidal galaxies in the GOODS-N
field 19 with (unexpected) 24 µm detection (12
also detected in X-rays) MIR to NIR luminosity
ratios in GOODs objects (ltzgt 0.7) 10x higher
than in local early-types detected by IRAS
(Knapp et al. 1992)
No mass loss from evolved giant stars dominating
MIR
32
? Broad-band SEDs MIR in excess with respect to
expectations for elliptical galaxies In most
cases SED well fitted by evolved stellar
pop (reproducing opt/NIR) dusty torus heated by
AGN (reproducing MIR/FIR) (Fritz et al. 2006)
fit with starburst fit with
ellipticaltorus
33
Pozzi et al. 2007, AA, in press,
astro-ph/0704.0735 The bolometric output of
luminous obscured quasars The Spitzer perspective
8 obscured AGN candidates at high z selected from
hard X-ray survey (high X/O ratio)
? Spitzer observations (3.6- 24 µm)
? All detected by Spitzer Dust reprocessed by
torus And thermally re-emitted
34
Bolometric Luminosity 1045 - 1047 erg
s-1 Bolometric Correction 25 High Stellar
Mass (0.8 - 6) x 1011 M? BH Mass (0.2 - 2.5) x
109 M? Eddington Ratio L/Ledd lt 0.1 ? low
accretion rate phase
35
Gruppioni et al. 2007, ApJ, submittedBroad-band
SEDs of a Complete Spectroscopic Sample of MIR
Selected Sources at intermediate z
200 MIR sources with spectroscopic z
(0.1-1.5) Data from far-UV 0.13 µm (GALEX) to
far-IR 160 µm (Spitzer) ? SED-fitting
SED-class compared with spec class
QSO
Seyfert 2
Data fitted with 21 Template SEDs of IR
galaxies/AGN by Polletta et al. (2007)
36
Example of fits with Galaxy/AGN Templates
? SED-fitting able to identify AGN activity
in 50 of MIR sources (to 0.5 mJy)
37
Normalised to 24 µm
Black Sd galaxy Green Seyfert 2 Magenta
starburst
Normalised to opt
Seyfert 2 warmer NIR/MIR
38
SED-fitting versus spectroscopic classification
SED Classification
Spectroscopic
Class S0/Sa/Sb/Sc/Sd/Sdm
Galaxy (83) Galaxy
(106) M82/NGC6090/Arp220
Starburst (13) Starburst
(32) Sey2/Sey1.8/RedQSO/I19254 AGN 2
(73) Liner/AGN 2
(31) QSO/Mrk231
AGN 1 (31) AGN 1 (25)

• Main differences
• SEDs classify more AGN (especially type 2)
• 2. Many galaxies with only Ha or OII emission
  in their spectra e(a) do show AGN-like SEDs
  

39
STRONG IMPACT ON EVOLUTIONARY MODELS?
? AGN/galaxy relative fractions
AGN fraction with new classification based on
SEDs (UPPER LIMIT?)
Matute et al. 2006 model predictions (LOWER
LIMIT?)
See also Brand et al. 2006 for a 24 µm selected
sample
40
YES STRONG IMPACT ON EVOLUTIONARY MODELS!
? Observed Source Counts
Spectroscopic classification
SED classification
AGN lower limit?
AGN upper limit?
GAL
TOT AGN
TOT
AGN 2
AGN 1
41
MAIN MODEL CHANGES

• Less objects powered by pure star-formation
• ? Less evolution for galaxies (starburst)
• SEDs evolving with LIR (or z) for all
• ? from S0 to Sd/Sdm for normal galaxies
• ? from M82 to ULIG for starbursts
• More objects containing an AGN
• ? More evolution for AGN (type 2)
• both luminosity and density (more numerous)?
• ? Similar evolution for type 1 AGN (slightly
• more objects)

42
need for HERSCHEL (3.5-m)?
Evolution of dusty galaxies up to IRAS ?
z0.2 ISOCAM ? z1.5 SPITZER ? z3
HERSCHEL will operate in the
FIR/sub-mm between 75 and 500 µm ? SF up to z4-5
43
need for HERSCHEL?
Locate the peak of dust emission in High-z
AGN/galaxies ? dust Temperature
detailed SEDs in FIR
z2
? Better Galaxy/AGN Separation!
Obscured AGN
Extreme Starburst
44
(No Transcript)