2011A&A...529A..72F

Nearly 100 new strong-lens candidates have been discovered in the COSMOS field. Among these, 20 lens candidates with 0.34≲z_lens≲1.13 feature multiple images of background sources. Using the multi-wavelength coverage of the field and its spectroscopic follow-up, we characterize the evolution with redshift of the environment and of the dark-matter (DM) fraction of the lens galaxies.We present spectroscopic and new photometric redshifts of the strong-lens candidates. The lens environment is characterized in the following way: we account for the projected 10 closest galaxies around each lens and for galaxies with a projected distance less than 1 Mpc at the lens galaxy redshift. In both cases, we perform similar measurements on a control sample of ``twin'' non-lens early-type galaxies (ETGs). In addition, we identify group members and field galaxies in the X-ray and optical catalogs of galaxy groups and clusters. From those catalogs, we measure the external shear contribution of the groups/clusters surrounding the lens galaxies. The systems are then modeled using a singular isothermal ellipsoid for the lens galaxies plus the external shear produced by the groups/clusters. We observe that the average stellar mass of lens galaxies increases with redshift. In addition, we measure that the environment of lens galaxies is compatible with that of the twins over the whole redshift range tested here. During the lens modeling, we notice that when let free, the external shear points in a direction which is the mean direction of the external shear produced by the groups/clusters and of the closest galaxy to the lens. We also notice that the DM fraction of the lens galaxies measured within the Einstein radius significantly decreases as the redshift increases. Given these observations, we conclude that while the environment of lens galaxies is compatible with that of non-lens ETGS over a wide range of redshifts, their mass properties evolves significantly with redshift: it is still not clear whether this advocates in favor of a stronger lensing bias toward massive objects at high redshift or if it is simply representative of the high proportion of massive and high stellar density galaxies at high redshift.