Introduction Active Galactic Nuclei
Lecture -10- Quasar Surveys and Evolution
Lecture at 9.00-10.30 on Tue June 17
th?
This Lecture
Give a general overview of quasar surveys and their LF evolution.
Read Chapt.10-11 of Peterson
(somewhat out of date, but read anyway)
● Techniques for finding quasars:
➔ Broad-band colors (nowadays)
➔ Slitless spectrosopy
● Some major surveys:
➔ e.g. SDSS, 2dF, Combo-17
● Introduction to the evolution of quasars (e.g. LFs)
Outline
Why do Quasar Surveys?
● Uniformly selected quasar samples form the input for studies of:
● Cosmology (less important nowadays)
● Quasar Structure & Evolution
● Co-evolution of AGNs, Galaxies & SMBHs
Measured Function: dN(F,z, {P})/(dF dz dΩ d{P})
(Number of AGN a.f.o. flux, redshift and area + rest)
-> From this a Luminosity Function can be derived a.f.o. Redshift -> Evolution?
Biases in Quasar Surveys
Flux-limit samples can be biased (mostly near the flux-limit) because of noise: Eddington Bias
● Because of noise (eg Gaussian), the observed
A(s) = dN(S)/dS is convolved with the noise PDF.
i.e. A'(S) ∝ A(s) ⊗ PDF.
● If dN(S)/dS steepens near the flux-limit, MORE faint sources will go into the sample, than brighter sources out of the sample -> #-count is biased!
Am≈ A' m−m2 2
d2 A' m
dm2
In magnitudes (see Peterson for derivation)
Biases in Quasar Surveys
More possible selection issues/biases:
● Variability:
Sources “fluctuate” in/out of the sample. In case of Gaussian distributed variability the effect is like the Eddington Bias.
● EW of emissions line vary:
Spectroscopic Surveys can be biased
● Absorption Lines:
Absorption by intervening IGM can alter spectra (in particular shortward of Lya)
● Internal Absorption:
Dust (e.g. torus) in AGN can obscure them (Type-II quasars?)
● The problem: AGNs are rare objects!
● How to find them among the much more numerous stars and galaxies on images of the sky?
● The solution: Use the observed ways that quasars/AGNs are “not stars.”
-> Make use of SEDs, emission-line spectra, variability, or lack of proper motions to distinguish AGNs from stars & galaxies
Finding Quasars and AGN
Crowed fields make AGN surveys difficult
Compact AGN and stars are
hard to separate using only a single color
● Consider the radio, optical/UV, and X-ray emission from AGNs
● Normal stars and galaxies emit very weakly in radio and X-ray regions.
● Normal stars and galaxies in general have “red”
colors at optical wavelengths
● How to use these properties in practice to find quasars/AGNs?
Use Spectral Energy Distribution
● Radio observations
– Led to the discovery of the first quasar
– Isolate a very important component of the quasar/AGN population
● Requirements for success
– Radio surveys to faint flux limits – Wide sky coverage
– Excellent positions
– Corresponding optical imaging &
spectroscopy
Begin with radio surveys
(previous lecture)
But not all AGN (only ~10%) are radio loud -> subsample
● The various surveys of all types need follow-up observations to confirm which objects are quasars/AGNs
– Generally, this means optical spectroscopy to establish that candidates are indeed AGNs and to determine their redshifts
Follow-up Observations
● Surveys should now be quantifiable in their
selection parameters (statistical completeness)
– selection efficiency as f(brightness, wavelength)
● Surveys also need to be effective
– efficient (high succes rate)
– but not overwhelmed with false positives, i.e. objects which meet criteria but are not AGNs (e.g. stars)
● This is all now possible with digitally based surveys
(old surveys were based on e.g. photographic plates)
Additional Survey Requirements
● Quasars/AGNshave strong, broad emission lines, different from emission regions ionized by stars
– This led Smith (1975) and Osmer & Smith
(1976) to develop the slitless spectrum technique:
– Very effective for finding quasars at z > 2 via direct detection of Lya emission
– Schmidt, Schneider, Gunn 1986; deep digital survey with slitless spectra
Optical Surveys:
Spectroscopic Techniques
Direct image (HST)
Slitless spectra
Example of Slitless Spectroscopy
● Historically:
– Quasars are more bright in the ultraviolet than most stars
– This allowed quasars to be found by their “UV excess”
– Very effective for redshifts z =2.3, but not higher, as Lya moves into the UV passband – At higher redshifts, quasars differ from
stars/galaxies at red and near-IR wavelengths
Optical Surveys:
Multi-color Techniques
Optical Surveys:
Multi-color Techniques
SDSS (Fan et al .1999) Quasar selection
Simulated Quasar color track
Multi-color observation allow AGN and stars to
be separated.
Optical Surveys:
Multi-color Techniques
At very high redshifts, contamination with
L dwarfs (very low-mass stars) might become a problem (Fan et al. 2001)
SDSS Quasar Spectra
SDSS Highest-z Quasar Spectra
Finding high-z quasars means observing at longer wavelenths!
Gunn-Peterson trough
SDSS Quasar Luminosity Function
Strong evolution with time!
● 2dF, UVX survey: >25,000 quasars, www.2dfquasar.org/
● SDSS, Multicolor survey: 77,429 quasars (2007, DR5; en route to 100,000)
www.sdss.org/
● COMBO-17: 300 quasars to R < 24,
www.mpia.de/COMBO/combo_index.html
Some Current Optical Surveys
Optical Surveys:
Multi-color Techniques
Combo-17 Survey: 12 narrow-band + 4 broad-band filters
Excellent photo-z determinations!
Optical Surveys:
Multi-color Techniques
COMBO-17:
Narrow-band filters are sensitive to strong
emission lines of AGN
Optical Surveys:
Multi-color Techniques
Optical Surveys:
Multi-color Techniques
COMBO-17: Resulting LF, note the strong evolution with redshift (peaking around z~2)
● X-Ray emission is perhaps the most common and distinctive feature of AGNs & quasars
– Historically, the first satellites did not have the sensitivity and spatial resolution needed for
AGN surveys.
– Now, thanks to Chandra and XMM, deep X-Ray surveys yield the highest surface densities of all techniques.
X-ray Surveys
See review by Brandt & Hasinger, 2005, Annual Reviews for a good summary Examples of important surveys: –ROSAT –ASCA –BeppoSAX –Chandra Deep Fields –XMM-Newton
Chandra Deep Field South
X-ray surveys allow detection of distant
X-ray quasars, also those shrouded in
gas and dust (e.g.
type-II quasars).
Chandra Deep Field North
HDF-North CDF-North
Identification in comparison to deep optical images
X-ray variability of AGN
At least 50% of the sources are variable, and that the fraction of variable AGNs may be be as large as 95%
(bias?)
(also some statistical fluctuations)
XMM Lockman Hole/Deep-Field North
Less resolution, but deeper and
X-ray spectra.
DFN
Chandra Number Counts
X-ray surveys provide excellent, deep point-source #counts of AGN (which dominate surveys out of the Galactic plane)
A bit more on AGN evolution
Evolution of quasar densities
Exponential decline of quasar density at high redshift,
different from normal galaxies, mostly luminosity dependent
Richards et al. 2005, Fan et al. 2005
SFR of galaxies Density of quasars
Bouwens et al.
Quasar Density at z~6
• From SDSS i-dropout survey – Density declines by a
factor of ~40 from between z~2.5 and z~6
• Cosmological implication – MBH~109-10 Msun
– Mhalo ~ 1012-13 Msun
– rare, 5-6 sigma peaks at z~6 (density of 1 per Gpc3)
• Assembly of massive dark matter halo environment?
• Assembly of supermassive
BHs? Fan et al. 2004
Evolution of the Shape of Quasar LF
Richards et al. 2005
High-z quasar LF different from low-z
– Bright-end slope of QLF is a strong function of redshift
– Transition at z~2-3 (where quasar density peaks in the universe)
– Different formation mechanism at low and high-z?
Richards, et al.; Fan et al. 2005
Evolution of the Shape of
Quasar LF
Simulating z~6 Quasars
• The largest halo in Millennium simulation (500 Mpc cube) at z=6.2
– Virial mass 5x1012 M_sun – Stellar mass 5x1010 M_sun – SFR: 300 M_sun/year
– Resembles properties of SDSS quasars
– Even the largest N-body simulation not big enough to produce one SDSS z~6 quasar…
– Today: 1.5 x 1015 M_sun cluster
– Much massive halos existed at z~6, but..
• How to assemble such mass BHs and their host galaxies in less than 1 Gyr??
– The universe was ~20 tedd old
– Initial assembly from seed BH at z>>10
– Little or no feedback to stop BH/galaxy growth
z=6.2
z=0
Dark matter galaxy
Springel et al. 2005
NV
OI SiIV
Ly a
Ly a forest
● Rapid chemical enrichment in quasar vicinity
● Quasar env has supersolar metallicity -- metal lines, CO, dust etc.
● High-z quasars and their environments mature early on
The Lack of Evolution in
Quasar Intrinsic Spectral Properties
Early Growth of Supermassive Black Holes
Vestergaard 2004 Dietrich and Hamann 2004
Billion solar mass BH at z~6 indicates very early growth of BHs in the Universe
Formation timescale (assuming Eddington)
Lack of spectral evolution in high-redshift quasars -> quasar BH estimate valid at high-z BH mass estimate: using emission line width to approximate gravitational velocity,
accurate to a factor of 3 – 5 locally
Summary
● AGN surveys are now dominated by a few large surveys like SDSS, 2dF (only ~10% of AGN are radio-loud)
● X-ray surveys are now an important new way of finding (also “hidden”) AGN
● The QSO LF evolves rapidly from z=0, peaks at z~2 and declines at z>2. But QSO properties (e.g. Spectra) and metalicity remain similar from z=0 to z~6.
● High-z AGN implies rapid SMBH growth in the early Universe