The spatial clustering of X-ray selected AGN

1
The spatial clustering of X-ray selected AGN

• R. Gilli
• Istituto Nazionale di Astrofisica (INAF)
• Osservatorio Astronomico di Bologna
• On behalf of the CDFS and COSMOS teams
  

2
The Large Scale Structure of the Universe
DARK MATTER
(theory)
12.6 billion yr ago z5.7
From Springel et al. 2005
9 billion yr ago z1.4
Now z0
GALAXIES
(observations)
(SDSS, Zehavi et al. 2002)
Matter distribution is everything but random
3

• AGN clustering
• AGN can trace the Large Scale Structure of the
  Universe
• to cosmologically significant redshifts (zgt1-2)
• AGN physics from clustering e.g. host halo mass
  and lifetimes
• Also used to estimate the host galaxy type
• in the X-ray band
• X-rays efficiently sample obscured AGNs (which
  outnumber
• unobscured ones by a factor of at least 3-4
  see XRB synthesis
• and the results from the deep surveys)
• High surface density, above 1000-2000
  sources/deg2, even in moderately deep surveys
• ? high clustering significance with relatively
  small number of objects

e.g. Croom 05, Porciani 05 Martini
Weinberg 01
4

AGN clustering as a function of.. 1)
redshift measured for optical QSOs r0 increases
with redshift ? halo mass constant 3 1012
Msun (e.g 2dF Croom et al. 04,05 Porciani et
al. 05,06) 2) luminosity still to be
measured 3) obscuration still to be
measured From optical and X-ray studies (Yang
05) there is no clear evidence for any dependence
of r0 on L. Similarly, there is no evidence yet
for any dependence of r0 on obscuration
(X-ray samples still small)
5
X-ray fields with measured 3D clustering

• Field Area/deg2 Slimcgs N/deg2
  N(zspec) zmed logLmed Ref.
• CDFS 0.1 5 10-17
  6000 130 0.8 43.0 Gilli 05
• CDFN 0.13 2 10-17
  8000 240 0.9 43.2
• CLASXS 0.4 5 10-16
  1500 230 1.2 43.9 Yang 05
• COSMOS 2 7 10-16 1000 380 1.0 44.1 in
  prep.
• NEP 81 2 10-14
  10 220 0.4 44.7 Mullis 03

Surveys with different sensitivities sample
different z and Lx regimes one needs to always
specify at which average z and Lx clustering is
computed
6
Statistical description of LSSthe 2-point
correlation function
Spatial Correlation Function excess probability
over random of finding a source in dV1 and
another in dV2 separated by a distance r

Projected correlation function w(rp)
Allows to get rid of distortions in redshift
space (i.e. peculiar velocities, redshift errors)
?(r) increases at small separations, then one
needs either deep fields in which the source
density is very high or large volumes or,
preferably, both
7
Expected 3D clustering significance in X-ray
surveys
SNR(DD-RR)/RR0.5 Where DD and RR are
the numbers of data-data and random-random pairs
within a given separation (e.g. 10 Mpc)
SNR V-0.5 N
8
3D correlation function in the Chandra Msec
fields
CDFS vs CDFN AGN
CDFN abs vs unabs AGN
Large variance between the two Chandra Msec fields
No statistically significant difference between
clustering of obscured and unobscured AGN within
a single CDF
9
Comparing redshift distributions
CDFS
CDFN
13
30 of sources in the 2 most prominent spikes
CLASXS
COSMOS
6
5-8
Due to redshift spikes, cannot divide in zbins
in the CDF-S/N Need large area to avoid cosmic
variance
10
XMM-COSMOS

CDF-S/N
0.1 deg2
CLASXS
2 deg2
0.4 deg2
First XMM-COSMOS results Hasinger 06, Cappelluti
06, Brusa 06, Mainieri 06, Miyaji 06, Finoguenov
06
11
Preliminary measurements in COSMOS
All sources
unobscured obscured
r0 7.3 0.5 ? 1.8 0.1
Based on Hardness Ratios
12
Clustering of obscured vs unobscured AGN
unobscured obscured
No statistical evidence for any difference. Based
on hardness-ratios need to be done using NH
measurements from spectral fits
13
AGN Correlation length vs redshift(and
luminosity)
Halo catalogs from Kauffman et al. 99

Late,early type r0 from Coil et al. 03
14
Summary

• Spatial clustering of X-ray selected AGN has been
  measured so far in only 5 fields. Large cosmic
  variance is observed between the two pencil beam
  Chandra Msec fields. Need large areas to get an
  unbiased estimate of AGN clustering and need
  large
• area x Nsources product (i.e. large and
  sensitive surveys) to measure clustering signal
  at high significance.
• Within each field no significant difference is
  found between the clustering properties of
  obscured and unobscured AGN (but check
• using X-ray spectra rather than HR).
• X-ray selected AGN appear hosted by halos with
  mass larger
• than 1-2 x 1012 Msun
• At z1 the AGN correlation length is in between
  that of early and late type galaxies at the same
  redshift (but large scatter COSMOS early).