Cooling and heating the intracluster medium in hydrodynamical simulations

We discuss tree+SPH (smoothed-particle hydrodynamics) simulations of galaxy clusters and groups, aimed at studying the effect of cooling and non-gravitational heating on observable properties of the intracluster medium (ICM). We simulate at high resolution four group- and cluster-sized haloes, with virial masses in the range (0.2-4) x 10(14) M., extracted from a cosmological simulation of a flat Lambda-cold dark matter model. We discuss the effects of using different SPH implementations and show that high resolution is mandatory to correctly follow the cooling pattern of the ICM. Our recipes for non-gravitational heating release energy to the gas either in an impulsive way, at some heating redshift, or by modulating the heating as a function of redshift according to the star formation history predicted by a semi-analytic model of galaxy formation. Our simulations demonstrate that cooling and non-gravitational heating exhibit a rather complex interplay in determining the properties of the ICM: results on the amount of star formation and on the X-ray properties are sensitive not only to the amount of heating energy, but also depend on the redshift at which it is assigned to gas particles. All of our heating schemes that correctly reproduce the X-ray scaling properties of clusters and groups do not succeed in reducing the fraction of collapsed gas below a level of 20 (30) per cent at the cluster (group) scale, which appears to be in excess of observational constraints. Finally, gas compression in cooling cluster regions causes an increase of the temperature and a steepening of the temperature profiles, independent of the presence of non-gravitational heating processes. This is inconsistent with recent observational evidence for a decrease of gas temperature towards the centre of relaxed clusters. Provided these discrepancies persist even for a more refined modelling of energy feedback from supernova or active galactic nuclei, they may indicate that some basic physical process is still missing in hydrodynamical simulations.