2014MNRAS.444.3581R

The metal content of clusters of galaxies and its relation to their stellar content are revisited making use of a cluster sample for which all four basic parameters are homogeneously measured within consistent radii, namely core-excised mass-weighted metallicity plus total, stellar and intracluster medium (ICM) masses. For clusters of total mass M500 ≃ 10¹⁴M_☉, nice agreement is found between their iron content and what is expected from empirical supernova yields. For the same clusters, there also appears to be at least as much iron in the ICM as there is still locked into stars (i.e. the ICM/stars' metal share is about unity). However, for more massive clusters, the stellar mass fraction appears to drop substantially without being accompanied by a drop in the ICM metallicity, thus generating a major tension with the nucleosynthesis expectation and inflating the metal share to extremely high values (up to ∼ 6). Various possible solutions of this conundrum are discussed, but are all considered either astrophysically implausible, or lacking an independent observational support. For this reason, we still entertain the possibility that even some of the best cluster data may be faulty, though we are not able to identify any obvious bias. Finally, based on the stellar mass-metallicity relation for local galaxies, we estimate the contribution of galaxies to the ICM enrichment as a function of their mass, concluding that even the most massive galaxies must have lost a major fraction of the metals they have produced.