Probing the Limits of Nuclear Activity with the COSMOS Survey

1
Probing the Limits of Nuclear Activity with the
COSMOS Survey

• Chris Impey (University of Arizona)
• Ph.D. Student Jon Trump

Collaborators Marcella Brusa (Max-Planck),
Martin Elvis (CfA), Pat McCarthy (Carnegie), Mara
Salvato Nick Scoville (Caltech), Yoshi
Taniguchi (Ehime), with help from the
international COSMOS Team
2
Active Galactic Nuclei (AGN)

• Nuclear activity in galaxies known since 1940s
• Luminous AGN (QSOs) discovered in 1960s
• Luminosity requires a gravity engine accretion
  onto a super-massive black hole (SMBH)
• QSO high luminosity, Seyfert low luminosity

Emission lines redshifted to z 0.16
Unresolved point source, like a star
3
Supermassive Black Holes

• SMBH Ubiquitous
• Cycles of activity
• Active, like AGN
• Passive, like Sag A in our Milky Way
• May be obscured by gas/dust
• From host galaxy?
• Or local to AGN?
• Key role in galaxy evolution

4
COSMOS Science Goals

• Tracing the coupled evolution of LSS, galaxies,
  SF, AGN
• Modeling LSS
• weak lensing for dark matter maps
• compare LSS and SF with simulation
• Similar volume to SDSS, but at z 1

5
SMBH and the Host Galaxy

• Tight relation between MBH and Mbulge or Lbulge
• Suggests that SMBH the host are connected

• Host controls the SMBH accretion?
• While SMBH regulates the star formation in the
  host galaxy?

from Ferrarese et al (2006)
6
(No Transcript)
7
HST Surveys
sizes of HST surveys
COSMOS
PI Nick Scoville
575 orbits covering 2 sq deg (10 of HST for 2
years) COSMOS is 9x any previous HST image 100
gigapixels Two million galaxies at z 0.2 to 5
(like Sloan, but high z)
8
6x zoom
Optical
Subaru 8m, PI Taniguchi 35 nights
Very deep images in 34 opt IR bands give colors
and redshift of galaxy, and morphology
9
X-ray
XMM, PI Hasinger 2.1 Msec
Sensitive to hot cluster gas, most efficient
way to find AGN, 1500 deg-2.
10
8x zoom
Infrared
Spitzer IRAC, PI Sanders, 600 hrs w/ MIPS
Cool dust, SF rate in galaxies, obscured AGN pops.
11
Radio
VLA, PI Schinnerer 300 hrs, to 7-10 ?Jy
Proxy for, SF rate in galaxies, radio-loud AGN
pops.
12
COSMOS Multi-? Photometry
5? in a 3 arsec aperture
Capak et al 2008
13
COSMOS 30-Band Photo-zs
Bright (lt22.5)
X-ray AGNs
Faint (lt24.5)
Spec-z VLT (Lilly et al), Keck (Capak et al),
Magellan (Trump et al)
?z / (1z) 0.7-1.6 to z 3 Galaxies
(Ilbert et al), AGN (Salvato et al)
14
Redshift Tomography 0.15 lt z lt 1.5
Stellar Mass
Galaxy Density
15
Stellar Mass
Galaxy Density
16
AGN Unified Model

• Orientation explains
• QSO/Seyfert
• Obscuration
• Type 1 (Broad Line) / Type 2 (Narrow Line)
• Radio emission
• But some models (e.g., Hopkins et al 2006)
  produce all these without orientation
• Observations give mixed support to the model

from Urry Padovani (1995)
17
Key AGN Questions

• What sets the episodes of activity?
• Fueling source
• Merger (Taniguchi 1999, di Matteo et al. 2005)
• Stochastic disk (Hopkins Hernquist 2006)
• Accretion rate
• Black hole mass
• What causes AGN obscuration?
• Geometry
• Host galaxy
• Accretion rate / Luminosity / Mass

18
COSMOS AGN Survey

• Faint
• IMACS spectroscopy is 4 mags fainter than main
  SDSS sample
• Smaller Black Holes and/or lower accretion rates
• Obscured
• X-ray, radio, and IR selection
• Obscured evolution to z1
• Rare populations
• X-ray bright, optically normal galaxies, BAL
  QSOs, high z
• Complete SEDs
• Bolometric luminosities
• Higher Redshift
• Host morphologies, environments obscured types
  to z 1

19
COSMOS
RA 100028.6 Dec 021221
20
COSMOS Multiwavelength

• Photometry
• VLA 1.4 GHz (Schinnerer) 7 µJy
• Spitzer-IRAC 3-8 µm (Sanders) 10 µJy
• Spitzer-MIPS 24 µm (Sanders) 15 mJy
• HST-ACS (Scoville) iAB27
• Subaru BVriz (Taniguchi) mAB27, 20 narrow
  bands to mAB 26
• GALEX N/F UV (Schiminovich) mAB26
• XMM (Hasinger) 0.5-10 keV 8x1016 cgs
• Chandra (Elvis) 0.5-8 keV 2x1016 cgs (new)
• Spectroscopy
• VLT/VIMOS (Lilly) 10,000 galaxies to iABlt26
• Magellan/IMACS (Trump/Impey) 1000 AGN to iABlt23

21
Sensitivity to AGN SEDs

• 40 times fainter than the typical SDSS quasar
• Sensitive to the QSO/Seyfert boundary at redshift
  2
• IRAC/MIPS for detecting the most heavily obscured
  AGN

SDSS SED, z1.5 (Richards et al. 2007)
Arp 220, z1.5
22
AGN by IR Selection

• Can select obscured Compton-thick targets
  missed by soft X-ray surveys

Daddi et al. (2007) and Treister et al. (2008)
show that heavily obscured AGN are very
common COSMOS reveals many more of the most
obscured AGN by IR depth
23
COSMOS AGN by X-rays
Keck
24
The Spectroscopy

• Magellan/IMACS to I lt 23 over 4 years

Supplementary MMT/Hectospec for additional blue
spectral coverage 1569 spectra, including 540
X-ray AGN
25
Magellan / IMACS

• Large 2520 field
• 200 - 450 slits / field
• 5600-9200 Å, 10 Å res.
• 29 nights, 2005-2008
• Nod shuffle technique

26
MMT / Hectospec

• Blue supplement to the IMACS data
• Wider wavelength range 3800-9200 Å, 3 Å
  resolution
• 1 diameter field
• 200 fibers / pointing
• Brighter limit (iABlt22)
• 4 nights, 2007-2008

27
Unobscured
Seyfert
QSO/Quasar
28
and Obscured AGN
Host SF-dominated
Type 2 AGN
XBONG
29
Measuring Black Hole Mass

• Type 1 AGN MBH L0.5 vfwhm2

• Calibrated from reverberation mapping of 30
  local AGN
• Virial theorem MBH RBLRvBLR2
• RBLRL0.5 (Kaspi et al. 2000, 07) scaling
  relations

30
Feeble Type 1 AGN
MBH L0.5 vfwhm2 (Vestergaard Peterson 2006,
2009)
182 Type 1 AGN from COSMOS Smaller black holes
than, e.g., SDSS zgt1 Sag A analogs
SDSS AGN (Kelly 08)
Typical error
31
AGN Fueling

• Type 1 AGN limited by L/LEdd gt 0.01

• L/LEdd increases with optical luminosity

32
AGN Fueling

• At L/LEddlt0.01, the BLR vanishes
• Obscured by clumpy torus (Nenkova et al. 08)
• Accretion disk changes (Hopkins et al. 08)
• L/LEdd increases with optical luminosity (4.8s
  significance)
• As L/LEdd, increases, more luminosity comes from
  the (cool) accretion disk
• At low L/LEdd, more luminosity is in (hot) X-rays

33
Obscured AGN Seen at zlt1

• L0.5-10 keV gt 3 x 1042 erg/s (Hornscheimer et al.
  2001, Alexander et al. 2001, Bauer et al. 2004,
  etc)

Type 1 (broad-line) AGN black cross Type 2
(narrow-line) AGN the blue diamonds XBONG
(absorption line only, host-light dominated)
the red squares
volume-limited complete zlt1 AGN sample
34
Unobscured to Obscured Ratio

• At z lt 1

Obscuration decreases with luminosity
(3.6s) Obscuration increases with redshift
2.4s) Fit using logistic regression
35
Obscuration with Luminosity

• Trim down the accretion disk
• Push out the dust sublimation radius
• Blow out the star-forming host gas
• Lawrence Elvis 1982, Simpson 2005, Ballantyne
  2008

36
Obscuration with Redshift

• The redshift evolution of obscured AGN matches
  the evolution of star formation
• Gas dust associated with star formation plays a
  key role in AGN obscuration!
• Barger et al. 2005,
• Ballantyne et al. 2008

Cosmic star formation history, (Hopkins 2004)
Increases to z1
37
Optically Dull AGN

• Bright AGN in X-rays, but high fX/fO ratio and no
  emission lines in the optical
• 50 in COSMOS 4 all previous samples
• Variability on year timescales for 3 objects

38
XBONGs

• Possible explanations for missing optical light
• Normal AGN, diluted by host (Caccianiga 07)

39
XBONGs

• Possible explanations for missing optical light
• Normal AGN in bright host (Caccianiga 07)
• Compton-thick nuclear absorption (Comastri 02)
• Compton-thin host absorption (Rigby 06)

Optically dull AGN lack the hot IR color of Type
1/2 AGN
IRAC color-color
Optically dull AGN are not Compton-thick
40
XBONGs

• Possible explanations for missing optical light
• Normal AGN in bright host (Caccianiga 07)
• Compton-thick nuclear absorption (Comastri 02)
• Compton-thin host absorption (Rigby 06)
• Truncated accretion disk (RIAF, Yuan Narayan 04)

41
XBONGs

• Explanations for missing optical light
• Normal AGN in bright host (Caccianiga 07)
• Works for optically dull AGN with normal fX/fO,
  70 of the COSMOS sample
• Nuclear absorption (Comastri 02, Civano 07)
• No evidence for hot IR dust
• Not Compton-thick
• Host galaxy absorption (Rigby 06)
• Variability on year timescales
• No evidence for edge-on preference of hosts
• Truncated accretion disk (Yuan Narayan 04)
• Explains 30 of optically dull AGN with high fX/fO

42
AGN hosts to z2
Stochastic disk accretion
Merger-fed
43
Bolometric Luminosity Function
Hopkins (2006)

• Bolometric luminosity can constrain the quasar
  lifetime, and the accretion mode. This can be
  done for the 1st time in a self-consistent way
  with the COSMOS survey, Jon Trumps thesis (late
  2009).

44
Narrow Band AGN Selection
Subaru narrow-band selection of Type 1 AGN
(Shunji Sasake Ph.D. Thesis)

• Select Type 1 AGN to iAB25
• Could easily be adapted to the selection of
  obscured X-ray Type 1 AGN
• Push studies of accretion and obscuration to
  fainter limits

45
Conclusions

• Sensitive to fainter, smaller AGN
• QSO/Seyfert boundary at z2
• MBH 107 Msun at z gt 1 detected
• Limiting accretion rate for a BLR
• Evolution of obscuration
• Bright AGN can blow out obscuration
• Significant obscuration from the host galaxy
• Still to come
• AGN Hosts and fueling constraints
• IR-selected sample, full obscured census
• Bolometric studies with deep multiwavelength data
• AGN from 30-band, IR to UV photo-zs