The phase diagram of the intergalactic medium and the entropy floor of groups and clusters: are clusters born warm?

We point out that two problems of observational cosmology, namely the facts (i) that ?60 per cent of the baryonic content of the Universe is not observed at z ∼ 0 and (ii) that the properties of small clusters do not agree with simple expectations, could be closely related. As shown by recent studies, the shock heating associated with the formation of large-scale structures heats the intergalactic medium (IGM) and leads to a 'warm IGM' component for the gas. In the same spirit, we suggest the intracluster medium (ICM) to be a mixture of galaxy-recycled, metal-enriched gas and intergalactic gas, shock heated by the collapsing much larger scales. This could be obtained through two processes: (1) the late infalling gas from the external warm IGM is efficiently mixed within the halo and brings some additional entropy, or (2) the shocks generated by larger non-linear scales are also present within clusters and can heat the ICM. We show that, if assumption (1) holds, the entropy brought by the warm IGM is sufficient to explain the observed properties of clusters, in particular the entropy floor and the L X -T relation. On the other hand, we note briefly that scenario (2) would require a stronger shock heating because of the larger density of the ICM as compared with filaments. Although the efficiency of these two processes remains to be checked on a quantitative level, they have the advantage of dispensing with the need to invoke any strong preheating from supernovae or quasars (which has otherwise been introduced for the sole purpose of reproducing the behaviour of clusters). Matter ejection by galaxies is included in the present calculations and, consistently with the metal-enrichment requirements, is indeed shown to yield only a quite moderate entropy increase. Our scenario of clusters being 'born warm' can be checked through the predicted redshift evolution of the entropy floor.