Cooling Flows in Clusters of Galaxies

1
Cooling Flows in Clusters of Galaxies

• A.C. Fabian ARAA 32, 277-318 (1994)
• and recent progress

2
The Origins of Cooling Flows

• Clusters discovered to be extended X-ray sources
  (Gursky et al. 1971, UHURU Kellogg et al. 1972
  Forman et al. 1972)
• Thermal emission was the natural interpretation
  (Felten et al. 1966 Lea et al 1973 Lea 1975)
  given the spectrum (Davidsen et al. 1975
  Gorenstein et al. 1973.) and extent.
• Intracluster Fe detected hot gas (Mitchell et
  al. 1976, Ariel V Serlemitsos et al. 1976 OSO8)
• Cooling times in many clusters very short. (Cowie
  Binney 1977 Fabian Nulsen 1977 Mathews
  Bregman 1978 Cowie, Fabian, Nulsen 1980)
• Unusual optical nebulae associated with cluster
  cooling flows. (Hu, Cowie Wang 1985 Heckman et
  al 1989Fabian, Nulsen Canizares 1984)

3
Some basic parameters of clusters

• Cluster mass 1014-1015 msolar
• X-ray luminosity 1043-1045 erg/s
• radius 1-2 Mpc, core radius 0.3-0.5 Mpc
• Intracluster gas temperature 107-108 K,
• number density 10-4-10-2cm-3
• total gas mass 51013-51014 Msolar (in rich
  cluster)
• chemical abundance Fe 0.3 times solar unit
• Cluster galaxies velocity dispersion 300-1200
  km/s
• Intracluster magnetic field 1-40 µG

4
Perseus ClusterNGC 1275 (Perseus A) Baade
Minkowski 1954
See also Hubble Humason 1931
5
Perseus Cluster WIYN (3.5m) Haimage Chris,
Conselice filamentary structure are clearly seen
6

• Image of the Perseus cluster from the Palomar Sky
  Survey. At the center is the cD (central
  dominant) galaxy Perseus A or NGC 1275

7

• High resolution of the central regions of NGC
  1275, where a smaller galaxy is being canabalised
  by the central galaxy. The central source and the
  filamentary structure are clearly seen. The large
  number of blue point sources are fresh clusters
  of stars forming from the shocked gas in the
  collision.

8
Zoomed in view of the very central regions of NGC
1275, showing the freshly formed clusters of
stars and dusty regions which are probably a
result of the collision. HST
9

• X-ray image of Perseus A taken with the Chandra
  telescope, showing a central bright source, two
  bubbles above and below this source, and the
  shadow of an infalling galaxy (top right). The
  remaining emission is from the central part of a
  cooling flow (the inward flow of hot gas into the
  potential well of the elliptical) which is losing
  energy (i.e. cooling) via thermal bremsstrahlung
  emission in the X-ray region.

10

• VLA radio image of Perseus A, overlayed on the
  X-ray image from the Chandra telescope. The
  central source is clearly seen, as well as radio
  lobes which apparently coincide with the two
  circum-nuclear bubbles in the X-ray image.

11
What is Cooling Flow?
Tgt3e7, thermal bremsstrahlung is the main
radiation mechanism.
12
Assumptions

• X-ray Luminosity is
• heat loss
• No heating
• Steady-state

Extra assumptions atomic physics determines L
and T, Locally Maxwellian, no absorption, metal
distribution, Exact prediction for mdot depends
on grav. potential
13
Evidence for cooling
The central luminosity is extraordinary high.
14
Radiative cooling times from Chandra
109yr
108yr
Other clusters Voigt Fabian 03
Perseus cluster
15
Evidence for cooling
Allen et al 01
The central part is cooler due to high number
density
16
X-ray spectrum low ionization emission line
4 actual cooling flows Mukai, Kinkhabwala,
Peterson, Kahn, Paerels 2003
17
Perseus Cluster
18
Non X-ray evidence for cool gas and young stars
Optical Crawford et al 99
UV Oergerle et al 01
CO Edge 02
Dust Edge et al 99
19
More than 30-50 of the clusters have surface
brightness tcoolltH0-1 within central 100
kpc. Cooling flow condition also occur in large,
isolated elliptical galaxies.
20
Evidence of a problem XMM/RGS
Sakelliou et al 02

• Gas drops to Tmin0.3Tvir
• Chandra spectra consistent
• (ltTmin)(0.1-0.2)

Peterson et al 01,02
McNamara et al David et al Allen et al
Blanton et al Buote et al
21
XMM spectroscopy

• Peterson, et al. 2003
• FeXVII and other lines from 1 keV gas not
  present.
• Two-temperature or truncated cooling flow (at
  T/3 - T/2)

22
Missing intermediate temperature X-ray lines
(OVII, FeXVII)

• FeXVII in Perseus
• then and now
• Einstein/FPCS 1 cm2
• XMM/RGS 60 cm2

23

• Result not explained by just heating gas at Tmin
  or just rearranging gas
• Since and
  at 1keV
• at
  constant P
• The cooler the gas, the faster it cools
• Why doesnt it cool?

24
The cooling flow problem (CFP)

• Why, and how, is the cooling of gas below Tvir/3
  suppressed?
• Significance
• The stellar/gaseous parts of galaxies are due to
  the cooling of gas in their DM potential wells
• Cooling in clusters and groups should be an
  observable example of this process.
• The suppression of cooling in these objects may
  explain the upper mass cutoff of galaxies.

25
Possible solutions of the CFP

• Absorption missing soft X-ray
• Mixing luminosity LUV
• Inhomogeneous metallicity
• Heating - from outside
• - from centre (AGN)
• - non-steady
• - combinations
• - other

26
Heat must be distributed
Voigt et al 03
27
AGN/Radio source

• Most CF have one, but no obvious correlation
• ? non-steady?
• Energy Budget
• seems inadequate for most systems (by 10x).
• It maybe not the dominant heat source

Voigt F 03

• Extreme energies required for luminous clusters,
  eg A1835, RXJ1347

28
Thermal Conduction

• Big Problem how is the conduction so smart.
• conduction coefficient needs to be 19 of
  KSpitzer for one cluster, 41 for another, etc.
• What is??

Voigt Fabian 03
29
Possible Solution add in the stellar mass loss
from stars in cD this gas is at 1 keV after
themalization (provides Tmin) the gas is flowing
outward into the cluster T will rise from Tmin to
Tcluster in all cases, but the shape varies with
?Spitzer Virtues not sensitive to ?Spitzer as
long as it is sufficiently large produces a
significant dT/dr that would be similar
from cluster to cluster dont get the
intermediate temperature cooling lines
30
New clues from deep Chandra observations of
Perseus A.C. Fabian et al., MNRAS. 344 (2003) L43
Three-colour image of the Perseus cluster from
Fabian et al (2000). Red indicates regions of
low-energy emission, blue indicates high-energy
emission. Visible in dark green above the centre
is an infalling dwarf galaxy, shown by absorption
of high-energy X-rays.
Chandra 200 ks observation of the Perseus
cluster.
31
The lowest note observed in the universe
Fluctuations in the Perseus cluster caused by
sound waves generated from the central black
hole. about 57 octaves lower than middle-C.
Temperature map of the Perseus cluster.
32
Ripples and weak shocks

• Bubbles make sound wave of long period (107
  years). Create weak shocks with some dissipation.
  
• Further out, dissipation depends on viscosity.
  What is viscosity?

Weak shock to NE
K108T 5/2n -1
Viscous heatingrad. cooling in inner 50 kpc of
Perseus
33
Collisional viscosity

• Has similar temperature dependence to conduction
• But MHD effects?
• Viscous dissipation could be crucial
• Turns sound energy from core and outer parts
  (refracted into core Pringle 89) into heat

34
Conclusions

• Cooling flow problem is widespread and plays a
  significant role in massive galaxy formation
• Many solutions proposed but all have drawbacks
• Knowledge of transport processes in ICM may be
  crucial to the solution

35
Cooling flow beyond Clusters
36
Interesting questions about clusters

• Why don't they cool as quickly as we expect?
• What's the history of the gas in the cluster, and
  how did it get filled by the chemical elements we
  see today?
• What is the cluster magnetic field role in the
  formation and evolution of clusters?
• How does the cluster interact with the central
  black hole?

37
There seems to be an association between these
events in a galaxy cluster CO emission Ha
emitting gas star formation AGN (radio lobe)
activity classic X-ray cooling flow
structure One interpretation of this is within
the Cooling Flow paradigm (at reduced
Mdot) Cooling Flow ? Cooled Gas (Ha CO) ?
star formation (possibly triggered by AGN)
38
Chemical evolution
39
Cluster magnetic field
Faraday rotation Centaurus cluster Chandra VLA
40
CF/B is correlated!
A correlation is found between the cooling flow
rate and the maximum Faraday rotation measures.
Magnetic fields of strength 1040 µG are found to
be common to the centres of clusters with strong
cooling flows, and somewhat lower field strengths
of 210 µG are found in the non cooling-flow
clusters. G. B. Taylor et al. MNRAS (2002)
334,769