The evolution of star formation properties of galaxies depends on the environment where galaxies reside, and generally star formation of galaxies in dense environment decreases more quickly. Interestingly, the star formation property of high-redshift galaxies clusters vary largely even though they are at similar redshift. We have found that the large-scale environment surrounding each galaxy cluster can contribute to make this cluster-by-cluster variation. This correlation is found in the results from observational data as well as in the simulations of galaxy formation. We suggest the ‘Web-feeding model’ to explain this trend. Star-forming galaxies falling into the galaxy cluster from surrounding large-scale structure make the quiescent galaxy fraction of the cluster lower than relatively isolated clusters.

In order to clarify the dust production in the early Universe, we constrain the dust mass in high-redshift (z ≳ 5) galaxies using the upper limits obtained by ALMA. We perform fitting to the rest-frame UV–far-infrared spectral energy distribution (SED) of a giant Lyα emitter, Himiko, at z = 6.6 and a composite SED of z > 5 Lyman break galaxies (LBGs). For Himiko, we obtain a high dust temperature > 70 K. This high dust temperature puts a strong upper limit on the total dust mass Md ≲ 2 × 106 M⊙, and the dust mass produced per supernova (SN) md,SN ≲ 0.1 M⊙. Such a low md,SN suggests significant loss of dust by reverse shock destruction or outflow. For the LBG sample, we only obtain an upper limit for md,SN as ∼2 M⊙. This clarifies the importance of observing UV-bright objects (like Himiko) to constrain the dust production by SNe.

Chandra revealed cavities in the hot atmospheres of many nearby clusters. These cavities are tracers of a strong coupling between the relativistic plasma in radio sources and the cooling, thermal gas in clusters. They demonstrate clearly that the AGN affects the cooling gas that leads to star formation and galaxy growth and allow a direct measurement of the bulk of the AGN's power. Together with radio data, the cavities allow us to derive scaling relations between mechanical (cavity) and radio power that can be used to estimate the AGN feedback power when direct measurement of the cavities is not possible. We review the importance of such relations for extending current studies of feedback with new and upcoming radio telescopes such as LOFAR and SKA.

Observations of CO molecules in the millimetrer domain at high redshift (larger than 1), have provided interesting informations about star formation efficiency, and its evolution with redshift. Due to the difficulty of the detections, selection effects are important. The detection if often due to gravitational amplification. Objects selected by their (far)infrared flux, are in general associated to ULIRGS, mergers with starburst in the nuclear regions. Quasars have been selected as powerful optical sources, and have been found to be associated to starbursts, rich in gas. The gas fraction appears to be much higher at redshift greater than 1. Quasars allow to probe the end of the reionisation period, and the relation between bulge and black hole mass. However these selection bias could have led us to miss some gaseous galaxies, with low-efficiency of star formation, such as the more quiescent objects selected by their BzK colors at z = 1.5 or 2.

Using the Very Large Array (VLA), we have detected the HCO+(1–0) emission line towards the Cloverleaf quasar (z = 2.56; Riechers et al. 2006). This is the first detection of ionized molecular gas emission at high redshift (z>2). HCO+ emission is a star formation indicator similar to HCN, tracing dense molecular hydrogen gas within star-forming molecular clouds. We find a HCO+/CO luminosity ratio of 0.08 and a HCO+/HCN luminosity ratio of 0.8 for the Cloverleaf. These ratios fall within the scatter of the same relationships found for low–z star–forming galaxies. However, a HCO+/HCN luminosity ratio close to unity would not be expected for the Cloverleaf if the recently suggested relation between this ratio and the far–infrared luminosity (Graciá–Carpio et al. 2006) were to hold. We conclude that a ratio between HCO+ and HCN luminosity close to 1 is likely due to the fact that the emission from both lines is optically thick and thermalized and emerges from dense regions of similar volumes. We conclude that HCO+ is potentially a good tracer for dense molecular gas at high redshift.