how to start an agn: the role of host galaxy environment

How to start an AGN:

the role of host galaxy

environmentRachel Gilmour (ESO Chile & IfA, Edinburgh)

Philip Best (Edinburgh), Omar Almaini & Meghan Gray (Nottingham)

or stop

Why do some galaxies have AGN?

Gas -> black hole = AGN

Internal: size, morphology, star-formation

Historical: previous activity -- depletion,

feedback

External: mergers, close encounters, tidal

field, strangulation, ram-pressure stripping

No AGN

0.01%

30%

External effects on galaxies

Morphology: spirals -> S0s

Star-formation rate: high -> low

Q1 – Do the frequency and properties of AGN

depend on the external environment?

Q2 – Can this be explained by the changes in the

type of host galaxies?

Where are AGN found?

Miller '03, Wake '05

Kauffmann '04, Coldwell '02

Dressler '99

Eckart '05

Ruderman '05

Johnson '03Porciani '04

Ledlow & Owen '96

Reddy '04

Best '04

Best '02

Galaxy density / AGN clustering

Red

shif

t

~0.8

~0.1

Optical

Radio

X-ray

My projects

1. AGN in the A901/2 supercluster• Detailed study

• X-ray detected AGN

• Includes groups, clusters, filaments, field etc.

2. Statistical survey of AGN in > 100 galaxy clusters• X-ray detected AGN

• Find statistical excess of sources compared to “blank” field

• Split sample by cluster properties and redshift

A901

A901aA901b

A902

Filament

SW group

The A901/2 supercluster (z=0.17)

Optical data (from COMBO-17 team)• Deep R-band imaging

• 17-band photometric redshifts for ~18000 objects (m

R<24)

• 1240 supercluster galaxies found

• 282 supercluster spectra from 2dF

• Weak lensing map

Spitzer data• MIPS 24-micron sources

X-ray data• 12 ksec ROSAT images (HRI)

• 90 ksec XMM-Newton image

People: Meghan Gray, Chris Wolf + COMBO-17 team, Bell and Papovich, Andy Taylor.

A901

A901a

A902

A901b

SW groupz=0.5 cluster

Emission

A901/2: Finding the supercluster AGN

• Identify point sources Sources- wavelet detection on images from 3 cameras 150

- remove uncertain and extended sources 139



- remove uncertain and extended sources 139• Match with optical sources

- R-band, using likelihood ratios from 14000 random sources

- also Spitzer 24-micron data to resolve uncertainties

88






88• Determine supercluster membership

- COMBO-17 photometric redshifts and 2dF spectra 11

- Manual check for AGN contaminated sources 12





88• Determine supercluster membership

- COMBO-17 photometric redshifts and 2dF spectra 11

- Manual check for AGN contaminated sources 12• Find out which are AGN

- check for low-mass X-ray binaries using fx/fB 12

- check for star-forming galaxies using Lx, hardness ratios, fx/fR,

star-formation rates from OII lines and OIII/Hβ line ratios 11


A901/2: Finding the non-AGN

Aim: Compare the AGN environments with control samples of galaxies

which have:• no AGN• similar magnitudes• similar colours• where AGN could be detected

Method:• 100 samples of 66 galaxies• equal number of galaxies in

each 0.5 magnitude bin• exclude cluster centres • use K-S and Kuipers tests

A901/2: AGN host galaxies

RESULT 1 : All of the AGN

lie in galaxies with mR < 20

RESULT 2 : ~5% of bright

supercluster galaxies contain

X-ray detected AGN (~1%

optically detected)

RESULT 3 : Brighter galaxies

have equal luminosity or

fainter AGN (92%)

A901/2: Separating the Environments

Define by hand:

• Clusters (A901a, A901a,

A901b, A902)

• Groups

• Outskirts of large clusters and groups

• Filaments

Two parameter

separation for

environments:

(1.5' = 250 kpc)

Local density:

clusters & field

Local colour:

groups & edges

filaments & clusters

1.5'

A901/2: Separating

the Environments - 2

A901/2: Separating the environments:

Does it work for all galaxies?

Cluster

Red group / Edge

Blue group

Filament

Field

cluster

filament

field

group

edge

A901/2: Environments of AGN

3.3% - AGN match

control in 2D

space

18% -AGN match

control in edge

and group

4% - AGN match

control in 1D

space along

cluster line (30%

in density only)

>98%-Lx

decreases along

the cluster line

direction

Cluster line

A901/2: Conclusions

1. ~5% of bright supercluster galaxies contain X-ray detected AGN.

2. All of the AGN lie in galaxies with mR < 20. more gas, larger black hole

3. The lack of AGN in fainter galaxies is not due to a LX – mR

correlation. no correlation or large galaxies are more stable

4. Compared to other similar galaxies, those with AGN lie in group or

edge like environments – moderate density and bluer than average.suppression in centre, triggering on outskirts, tracing star-formation, more smaller galaxies

5. AGN in more cluster like environments are fainter.galaxies with more gas need less disturbance, strangulation reduces available gas




Dressler '99

Eckart '05

Ruderman '05


Ledlow & Owen '96

Reddy '04

Best '04

Best '02


Red

shif

t

A901/2

~0.8

~0.1

Optical

Radio

X-ray

Chandra Clusters: Method

• Find sources in fields of galaxy clusters

• Predict source distribution assuming no cluster AGN

• Compare flux and radial distributions of excess sources

AGN

Abell 1689

HST Credit: NASA / N. Benitez

Chandra Clusters: The sample

Secure redshift and z > 0.1Exposure > 10 ksecX-ray detected cluster (after data reduction)

=== 139 good cluster fields ===

+ 8 with z > 1

Redshift distributionMorphology

89 'relaxed' 23 'disturbed' 19 'contaminated' 8 high redshift

(z > 1)

Luminosity

z

0.1 – 70 x 1044 erg/sec

Chandra Clusters: PredictionBlank fields – deep surveys (22) and high redshift QSOs (22)Sensitivity map – background, size, exposure, accuracy + errors

Background

Source size Good region

Exposure time

Sensitivity

map

Chandra Clusters: Lensing

Lensing changes background sources : flux increases

number density decreases

Net result: lensing causes ~ 10% reduction in the central 0.5 Mpc

INPUTS:

• Background AGN redshift distribution (3 used)

• Cluster model (SIS now, NFW in future)

• Cluster luminosity => mass

N

Lx

Model = Blank fields + Sensitivity map + Lensing

Chandra Clusters: Radial position

AGN lie between 0.5 and 1 Mpc

from the cluster centre.

Excess of 1 or 2

sources per cluster

Radial trend seen in

physical distance (Mpc)

Lack of AGN in central

regions is not due to

the intra-cluster

emission

Chandra Clusters: Suppression?

AGN appear to be suppressed

in moderate redshift clusters

NX(flux) NX(Cl,flux)

NOpt(>L*) NOpt(Cl,>L*) =

25 galaxies > L*

5 galaxies > L*

1 galaxy > L*

Clusters with low LX (~1x1044) have ~6 galaxies > L*

(De Propris 2004)

(Excess per average field = excess per square degree x average field size)

Chandra Clusters: Evolution

The evolution of AGN in clusters

is faster than in the field

25 gal.

25 gal.

25 gal.25 galaxies

5 galaxies

Redshift samples < 1 have

similar luminosity &

morphology distributions

Chandra Clusters: Radial Evolution

High redshift clusters have more

AGN at larger radii

Chandra Clusters: Morphology

Disturbed clusters have more

low luminosity sources

Disturbed clusters also have excess at higher radius

Chandra Clusters: Results

AGN lie between 0.5 and 1 Mpc from the cluster centre.

AGN appear to be suppressed in moderate redshift

clusters.

The evolution of AGN in clusters is faster than in the field.

High redshift clusters have more AGN at larger radii.

Disturbed clusters have more low luminosity sources.




Dressler '99

Eckart '05

Ruderman '05


Ledlow & Owen '96

Reddy '04

Best '04

Best '02


Red

shif

t

~0.8

~0.1

Optical

Radio

X-ray

A901/2



AGN appear to be suppressed in

moderate redshift clusters.


is faster than in the field.


AGN at larger radii.

Disturbed clusters have more low

luminosity sources.

Results~5% of bright supercluster

galaxies contain X-ray detected

AGN. All are in galaxies with mR < 20.

The lack of AGN in fainter galaxies is not due to a LX – mR

correlation.

Compared to similar galaxies,

those with AGN lie in group or

edge like environments –

moderate density and bluer.

AGN in more cluster like

environments are fainter.



AGN appear to be suppressed in

moderate redshift clusters.


is faster than in the field.


AGN at larger radii.

Disturbed clusters have more low

luminosity sources.

AGN in more massive clusters

have a larger radial spread.

Results~5% of bright supercluster

galaxies contain X-ray detected

AGN. All are in galaxies with mR < 20.

The lack of AGN in fainter galaxies is not due to a LX – mR

correlation.

Compared to similar galaxies,

those with AGN lie in group or

edge like environments –

moderate density and bluer.

AGN in more cluster like

environments are fainter.

Chandra Clusters: Cluster size

AGN in more massive clusters

have a larger radial spread

Expect mass to go as

Lx M4/3

Expect radius to go as

R ~Lx 0.3

how to start an agn: the role of host galaxy environment

Documents