Research: Galaxy Clusters

Clusters of galaxies are the largest gravitationally bound structures in dynamical equilibrium in the universe and as such are extremely interesting both in cosmological terms and in their own right. They consist mainly of dark matter and hot intracluster gas with a temperature of 10-100 million Kelvin, the galaxies themselves constituting a small percentage of the total mass of the cluster. The Birmingham group has programmes underway to study the hot gas and the galaxies in clusters using a combination of X-ray and optical observations coupled with modelling and dynamical calculations.

One important feature of galaxy clusters is that radiation from background sources, such as quasars or distant galaxies, is bent as it passes near the deep gravitational potential well of the cluster. This phenomenon, known as gravitational lensing, can be used to explore the cluster potential, the lensed sources and the geometry of the Universe.

On even larger scales, clusters are grouped into superclusters. These huge filamentary structures extend over tens of millions of light years, and the gravitational pull of their very large masses cause large scale patterns in the velocities of galaxies. For example, the Milky Way is being pulled towards a region called the "Great Attractor".

Further details of our work in these areas can be found below:



The Coma cluster is one of the nearest rich clusters of galaxies. It had generally been thought of as a rather structureless and well evolved cluster. The X-ray image above left, obtained with the XMM-Newton telescope, shows clear signs of complex substructures in the hot intracluster gas. This probably indicates that the cluster has a lively recent history, and has swallowed a number of smaller clusters within the past few billion years. For comparison, the scale bar in the top left corner is 1/3 of the angular size of the full moon. The optical image of the cluster centre on the right, shows the two yellowish giant elliptical galaxies which dominate the core region. Almost all the sources seen in this image, apart from the bright blue foreground star, are galaxies within the cluster.

Galaxy-Cluster interactions

Most galaxies in the nearby Universe reside in groups and clusters of galaxies. The dominant component of visible matter in these systems is hot, X-ray emitting gas. It is believed that interactions between group/cluster galaxies and their gaseous environment can strongly modify the properties and evolution of the galaxies themselves and the group/cluster as a whole. Various pieces of evidence exist today in support of the idea that galaxy-cluster interactions take place. For example:

  • In almost all cases, the jets of cluster radio galaxies show severe distortions compared to the well-collimated shapes of similar radio galaxies outside clusters.
  • The intracluster gas contains a vast amount of heavy elements (e.g., Fe and O). It is well-known that heavy elements are produced by stars inside galaxies. Galaxy-cluster interactions should play a major role in the removal of metal-rich gas from galaxies and distributing it throughout the cluster.
  • Many group and cluster galaxies show distorted gas morphologies, probably due to the pressure felt by the galaxies as they move through the cluster gas.

The study of galaxy-cluster interactions and their impact on the galaxies and clusters are a special interest of the group at Birmingham. There is information about some of our projects below.

Galaxy - Cluster interactions: Tailed radio galaxies

We are currently studying the interaction of radio jets with the wider cluster environment, specifically:

  • how the cluster medium affects radio galaxy morphology, e.g. the formation of wide angle tailed radio galaxies (Jetha et al 2005, Hardcastle & Sakelliou 2003).
  • how the cluster and the motion of the radio galaxy through the cluster causes jet bending.

To facilitate our program we have radio observations (VLA and ATCA) of the most powerful radio galaxies in clusters, XMM observations of merging clusters with tailed radio galaxies, and Chandra observations of tailed radio galaxies to investigate the regions where the jets decelerate.


A Wide-Angle Tailed Radio Galaxy (WAT) (e.g. 3C465 above) is a radio galaxy that contains very well collimated inner jets, which resemble an FRII radio galaxy (Fanaroff & Riley 1974), that flare suddenly into diffuse plumes which resemble those found in FRI galaxies. In addition, the jets and plumes can also be bent into wide C-shapes (e.g. 2151+085 below).


Using Chandra X-ray data for 0110+152 and 2230-176 we have found that the base of the radio lobes co-incides with a sharp increase in the temperature gradient (Jetha et al 2005). However, it is unclear as to whether the location of the increase in the temperature gradient is due to the presence of the radio lobes, or if the location of the lobe base is determined by the presence of the cool core. Further work is required to determine which of these scenarios is correct.

More radio images can be found here.

Galaxy-Cluster interactions: Gas stripping in galaxy groups

We are studying the physical processes through which galaxies are affected by their environment. It is well known that the hot, dense X-ray gas in clusters of galaxies can have an impact on the evolution of cluster galaxies. For example, galaxies travelling through this hot cluster gas will feel a "wind" which can strip out gas from within the galaxy. However, it is less clear whether this can also occur in smaller systems, groups of galaxies, where most galaxies live. We are investigating the importance of such processes in groups, by combining X-ray, optical, and radio observations of group galaxies.

The example shown below is a Chandra X-ray image of the hot gas in the spiral galaxy NGC 2276, a member of a small galaxy group. The gas in this galaxy appears compressed to the right of the galaxy, and stretches out in a tail in the opposite direction. Because of its motion through the hot gas in the group, this galaxy is being stripped of its gas at a rate of several solar masses per year.


We intend to follow up on such observations with detailed hydrodynamical simulations, to better constrain the nature and importance of the various physical processes associated with gas stripping from this galaxy, and in galaxy groups in general.

Researchers: Trevor Ponman, Ian Stevens, Jesper Rasmussen, Nazirah Jetha

Superclusters and large-scale filaments of galaxies

Superclusters are the largest gravitationally bound systems in the Universe. They are dynamically unrelaxed entities, and consequently tend to be more filamentary than clusters themselves. Using statistics sensitive to filamentary structures (e.g. minimal spanning trees), Somak Raychaudhury and collaborator Suketu Bhavsar (University of Kentucky, USA), have compiled a catalogue of superclusters from an almost complete redshift survey of Abell clusters out to z=0.1 (obtained by John Huchra and collaborators). This will help us to delineate the largest bound structures in the local Universe. Further research using this survey will include:

  • The clustering statistics of superclusters, which will reveal the clustering properties of matter in the linear regime, and the nature of dark matter haloes
  • Studies of individual superclusters combining optical observations with X-ray observations from ROSAT, ASCA and XMM


All clusters of galaxies with redshifts z < 0.2 in the Abell catalogue of clusters. The 739 clusters with z < 0.1 are plotted in blue, and the two richest known superclusters among them are marked. There are 2052 clusters between z=0.1 and 0.2. These clusters are optically selected, and many of their redshifts are estimated from the magnitudes of their few brightest galaxies. Note that there are very few clusters in the catalogue within 30 degrees of either side of the plane of the Milky Way.

We have been searching for diffuse Mpc-long radio emission (at 320 MHz and 1.4 GHz) from VLA and GMRT observations. A previously unknown (>6/h Mpc) filament (Zw2341.1+0000) has been discovered at z=0.3, for which X-ray, radio and optical data together suggest that it is a supercluster in formation. The energetics of accretion shocks generated in forming large-scale structures (giving rise to x-ray emitting gas) are sufficient to produce enough high energy cosmic-ray electrons that is required to explain the observed radio emission, provided a magnetic field of strength B > 0.3 microGauss is present in the intercluster medium.


The Pisces-Cetus supercluster, the most prominent supercluster in the 2dFGRS survey region, consisting of 25 clusters of galaxies, defined by a minimal spanning tree analysis of our nearby clusterS sample. Each cluster, represented as red circles, is at a distance of less than 20/h Mpc from its nearest neighbour (except when shown as dashed line). The axes represent Right Ascension and Declination (Porter and Raychaudhury 2004).

Researchers: Somak Raychaudhury, Scott Porter

Scaling Propeties of Groups and Clusters of Galaxies

Since gravity is the dominant force in the formation of galaxies and galaxy systems, the composition by mass of these objects is of great importance in understanding their behaviour. There are three main contributors to the mass of groups and clusters of galaxies, namely dark matter (80-85%); hot,  X-ray emitting gas (10-15%); and stars (2-5%).  The dominance of the dark matter component ensures that the properties of these systems are largely driven by gravity. Gravity is a scale-free force, and therefore the behaviour of a dark matter dominated halo is determined mostly by its mass. This implies that galaxy clusters of different masses resemble simple scaled versions of each other.  As a result, observable parameters such as the luminosity or gas temperature of these objects are expected to follow very predictable relationships, because these quantities reflect the depth of the cluster gravitational potential well (i.e. the total mass).

The following images show the X-rays emitted by the hot gas, as well as the corresponding optical view of the galaxies embedded in this gas. The first case is a large cluster of galaxies, and the panels below that show two typical galaxy groups. Despite being roughly 100 times less massive than the cluster, the groups of galaxies resemble a smaller-scaled version of it.

An X-ray/optical image of the Abell 2029 cluster

X-ray and optical images of the Abell 2029 galaxy cluster. (Credit: NASA)

X-ray/optical image of the NGC 5044 groupX-ray/optical image of the NGC 507 group
X-ray (blue colour table) and optical images of 2 galaxy groups. (Credit: NASA)

Numerical simulations of clusters show that the dark matter component behaves in a self-similar way, aside from slight variations between halos formed at different epochs, when the mean density of the Universe was different. However, most observations of real clusters don't generally measure the mass directly, but instead study the radiation emitted by luminous material which is confined by gravity, thereby tracing the mass.  Unlike dark matter, however, luminous matter (often referred to as "baryons") is subject to a range of complex physical processes, that alter the extent to which it tracks gravity. For this reason the scaling properties of real clusters are different from simple predictions. For example, the X-ray luminosities of poor clusters and groups are known to be significantly lower than expected, and such systems are found to contain proportionately much less hot gas in their inner regions than massive clusters.  Such behaviour points to the influence of physical processes originating within galaxies that can interact with and modify the hot gas (also known as the intracluster medium, or ICM) within the cluster or group. Examples of such feedback interactions include outbursts from active galactic nuclei (AGN) or supernove-driven galaxy winds, which can both heat and displace the ICM.

Since the energy injected by feedback mechanisms does not depend on the mass of a cluster, the impact it has will vary for different clusters. For instance, a large AGN outburst will have a very big effect on the ICM in a small group, where it is capable of displacing large amounts of material, effectively blowing the gas right out of the inner regions and dramatically reducing the amount of remaining gas. However, the same outburst occuring in a large cluster would have a much lesser impact, resulting in a smaller fractional decrease in the mass (and hence luminosity) of the gas near the AGN. This differential behaviour is revealed when the properties of clusters and groups of different masses are compared, across a large number of objects. Consequently the scaling properties of galaxy systems can be used to shed light on the nature of feedback mechanisms and their implications for galaxy formation and evolution.

Early analysis of cluster scaling relations was limited to simple global properties, such as the total X-ray luminosity of the hot gas and its mean temperature (expressed in units of keV). However, with better X-ray telescopes, it has been possible to study the ICM in more detail across a large number of systems.  The following two graphs show how the hot gas varies in temperature and density for a sample of 20 clusters observed with the Chandra X-ray telescope.  The temperatures range from massive clusters (blue), through small clusters (green) and down to groups of galaxies (red), with half the sample showing signs of strong cooling in their cores. The gas density varies considerably compared to the case of self-similarity, where the profiles would appear stacked on top of each other. Clusters with cool cores have denser gas than those without, and in general the density of the gas decreases for smaller (and hence cooler) clusters. This behaviour can be explained by the actions of galaxy feedback processess,  which have energised the gas and caused it to puff up more within groups, due to their weaker gravity.

Cluster gas temperature profiles    Cluster gas density profiles  
        Gas temperature and density profiles for a sample of 20 clusters, colour-coded according to their mean gas temperature.

The area of cluster scaling relations is a key su bject of ongoing research at Birmingham, and we are currently analysing a sample of galaxy clusters and groups observed with the Chandra and XMM-Newton X-ray telescopes, which provide a much more detailed view of the intracluster medium. This study will provide new insights into the close mutual interaction between galaxies and the intracluster medium surrounding them, which will lead to a better understanding of the formation and evolution of structure in the Universe.

Researchers:  Alastair Sanderson, Trevor Ponman, Rowan Temple, Paul Russell

X-ray Surveys: The XMMLSS survey for high redshift groups, clusters & AGN

By studying a large, continuous area of sky at faint X-ray fluxes we are able to easily locate distant clusters and groups of galaxies and active galactic nuclei up to high redshifts, and study their evolution with redshift and distribution in 3D. We are part of the XMMLSS consortium constructing this large, multiwavelength survey using data from radio wavelengths, through the near and far infrared, optical and ultraviolet to X-rays. In particular we will study the evolution of the hot gas and galaxies within groups of galaxies, very high redshift clusters, and fossil groups.

Researchers: Trevor Ponman, Abdulmonem Alshino, Somak Raychaudhury, Graham Smith


This X-ray image of the extragalactic sky covers one of the largest areas of sky at faint X-ray fluxes ever obtained. It contains approximately 600 X-ray sources in an area of 2 sq deg (equivalent to 10 times the full moon). This survey is being used to locate and map out the distribution of distant clusters and groups of galaxies, and QSOs. An example of an extended cluster of galaxies can be seen above and slightly left of the image centre. The two white circles denote the only sources detected in this area in the ROSAT All-Sky survey at bright fluxes (one was probably a false detection).


The REXCESS Large Project is a statistically unbiased sample of galaxy clusters selected from the REFLEX parent sample, observed with XMM-Newton. These were selected in luminosity-redshift space (Figure 1: for larger versions of all images, please open in new window), from which a homogeneous luminosity range was selected with no biasing on morphology. This sample was designed to fit inside the XMM-Newton field-of-view to assist with background statistics in the data analysis.


Figure 1. X-ray luminosity-redshift distribution of the REFLEX sample (small dots). The REXCESS sample are the data points selected in yellow boxes. The solid line is the survey flux limit, and the dashed lines are the distances at which R500 is 7, 9, 10, 12 arcmin respectively.

This cluster sample is relatively nearby (z < 0.2), which makes it an ideal sample for determining cluster statistics and scaling relations, which can be used for higher redshift Large Projects and cluster surveys.

This is a collaborative effort between many institutions and much of the work undertaken presently is to resolve differences between the different methods of data analysis. The work in Birmingham has recently revolved around two different types of analysis: morphology and mass.

  • Morphology. As this is an unbiased sample based on morphology it it provides an opportunity to study cluster irregularity and see if that correlates with other global gas properties of the cluster. A centroid shift method was used to quantify cluster asymmetry (Figures 2,3). A high morphology parameter implies a more irregular cluster than a system with a low morphology parameter.


    Figures 2 (left) and 3 (right): Morphological structure of RXCJ1302-0230. Blue circles are 0.3,0.5 R500. Contours are from smoothed X-ray image. The centroid shift of the cluster is indicated in the second figure off to the North-West.

    Cluster irregularity did not correlate with many cluster properties, with the exception of the entropy scaled by the empirically determined T^0.65 (Figure 4). The distribution of cool-core and non-cool core systems is indicated by the colour scale, and more irregular systems tended to have higher scaled entropy.

  • RFT_Fig4.jpg

    Figure 4: Scaled entropy (S/T^0.65) vs. Morphology. Red=non-cool-core clusters, Blue=cool-core clusters. Line=least squares best fit to data.

  • Mass.The X-ray total mass density can be derived under the assumption of hydrostatic equlibrium, and is dependent on radial temperature and density profiles. These were derived from an annular spectral analysis in XSPEC, with the appropriate profiles fitted parametrically. From this total mass profile, an NFW profile was fit to each cluster, to determine concentration parameters (Figure 5).


    Histogram of derived concentration parameters , c200. Red=non-cool-core clusters, Blue=cool-core clusters.

    The gas mass was subtracted off the total mass to leave mass density profiles that represented the stars and dark matter halo (Figure 6)


    Figure 6 (right): Dark matter + stars mass density profiles. Cluster irregularity is indicated by the colour bar (blue=more irregular). Line=NFW profile (arbitrary normalisation) using a concentration parameter of c200=4.

Researchers: R. Temple, A. Sanderson, T. Ponman (Birmingham)
External Researchers: H. Boehringer (PI), G. Pratt, A. Finoguenov (MPE, Garching), J. Croston (Hertfordshire), M. Arnaud (CEA,Saclay) + others in the collaboration