The simplest explanation for the common appearance of cold core, X-ray peaked clusters is that, when averaged over tens of Myr, the radiative cooling is balanced in part by distributed heating. Thermal conduction as a means of distributing heat from outer gas is ruled out for low and intermediate temperature clusters. It may however have a role in spreading the energy in the central parts. A plausible mechanism is the dissipation of energy propagating through the ICM from a central radio source. Such a process stems massive cooling onto the BCG which would otherwise gain a total stellar mass >> 1012 M. The process is therefore a vital ingredient in stopping the growth of the most massive galaxies (Fabian et al. 2002b, Benson et al. 2003, Binney 2004). Note that most semi-analytic models for galaxy formation (e.g. Kauffmann et al 1999) already needed to suppress cooling in massive haloes in order to match observation.
Difficulties and doubts remain with regard to the issues of the energy dissipation and distribution processes which are tied in with the transport processes in the gas. Similarly, it is not clear how the feedback manages to produce such similar cooling time profiles in systems where temperatures and thus masses differ by over an order of magnitude? There still remains the possibility that some process not yet foreseen, or at least not well studied will eventually prove more important than the effect of the central radio source, or will at least be important in mediating its effect. The central radio source is so common and so energetic however that it must at least be part of the solution. Similarly, the motions of galaxies or interacting dark matter, if it exists, could be important in heating cluster cores along with the AGN. Given the wide range of objects in which a balance is required, we suspect that a single mechanism is dominant, rather than several.
The need for a heating-cooling balance is in a time-averaged sense, over intervals of about 108 yr. In most cases the heating has not been so energetic as to drive gas out of the inner regions nor so weak as to allow much cooling at very high rates. Examples of objects at the extremes are Hydra A and Cygnus A where heating is high (but is dumped mostly at large radii outside the cool region) and A2597 and RXC1504.1-0248 where cooling appears to be high (in the latter object over 70 per cent of the total X-ray luminosity emerges from the cool core; Böhringer et al. 2005).
In most objects residual cooling at a rate of about 10 per cent of the simple unheated cooling rates appears to occur. It could be larger if non-radiative cooling, due to mixing say, is occurring. Stars form from the cooled gas giving the excess blue light seen in the BCGs. Mass loss from such stars can make the cooled gas dusty and radio bubbles drag some of it back out to large radii.
The time evolution of the heating / cooling balance is little understood. We suspect that the common temperature drop associated with the short central cooling times is due to radiative cooling and that heating only came into balance when the overall temperature structure was in place. Perhaps the central galaxy and its central BH grew until balance was achieved, and growth required cooling. Comparison of samples of clusters at = 0.22 with = 0.056 (Bauer et al. 2005) shows little surprisingly change in the distribution of cooling times at 50 kpc. The results imply that any balance was established well beyond z ~ 0.3.
Few massive cool core clusters are known at much higher redshifts than that of RXJ1347 at z = 0.44. This may however be a selection effect. They will be absent from X-ray cluster samples if the central BH is a bright X-ray source such as a quasar. If a central quasar outshines the host cluster in X-rays then the object will generally be classified as a quasar. H 1821+643 is a good example of a bright quasar in an X-ray peaked cluster at intermediate redshift (z = 0.297, Fang et al. 2002). Searches for clusters around powerful radio-loud quasars and galaxies have found some examples at z = 0.5-1.1 (Worrall et al. 2001, Crawford & Fabian 2003, Siemiginowska et al. 2005) but no complete searches have been done.