The analysis of a large sample of X-ray selected nearby galaxy clusters, allow us to perform an accurate determination of the fraction of baryon in stars as a function of cluster mass, to shows that clusters display a scatter (in the stellar, gas and baryon fraction), showing that the regions from which clusters and groups collect matter are not yet representative of the mean Universe content. Stellar and gas mass fractions at a given cluster mass are compared with theoretical predictions, showing that X-ray cluster properties can be matched to theoretical models for parameter choices in major disagreement with stellar fraction measurements. We also found that X-ray luminosity and richness (n200, the number of bright red galaxies) perform identically as mass proxy: both allow to predict mass with a 0.3 dex accuracy. However, richness is observationally easier to collect, with obvious cosmological consequences.

We performed an X-ray follow-up of an optically (n200) selected sample of local clusters, showing that all them are real systems, with potential deep enough to hot and retain an intracluster medium. From these observations we infer that at most 6 % (at 90 % confidece) of colour-selected clusters are fake systems, and that X-ray and colour searches sample the same cluster population and thus that we safely use clusters selected with any of the two methods for cosmological purposes.

Finally, we adress the measurement of the evolution of the ICM (Lx-T relation), by observing two, carefully selected, clusters at z about 2, hence maximizing the redshift leverage, finding them about 5 times less luminous than expected if evolution were self-similar in the last three quarter of the Universe age.