The population of early-type galaxies at 1 < z < 2 - new clues on their formation and evolution

We present the morphological analysis based on Hubble Space Telescope HST-NICMOS (Near-Infrared Camera and Multi-Object Spectrometer) observations in the F160W filter (lambda similar or equal to 1.6 mu m) of a sample of 32 early-type galaxies (ETGs) at 1 < z < 2 with spectroscopic confirmation of their redshift and spectral type. The 32 ETGs at < z > similar to 1.5 are placed on the (<mu >(e), R-e) plane according to the Kormendy relation (KR) with the same slope of the local one but with a different zero-point, which accounts for the evolution they undergo from z similar to 1.5-2 to z = 0. The best fitting of their spectral energy distribution shows that ETGs at 1 < z < 2 are composed of two distinct populations, an older population (oETGs) and a younger population (yETGs) whose mean ages differ by about 1.5-2 Gyr. Young ETGs are not denser than local ones since they follow the size-mass relation of local ETGs, and luminosity evolution brings them on to the local KR and size-luminosity relations without the need of size evolution. Old ETGs do not follow the size-mass relation of local ETGs, and luminosity evolution does not account for the discrepancy they show with respect to the local size-luminosity relation and KR. An increase in their Re by a factor of 2.5-3 (a density decrease by a factor of 15-30) from z similar to 1.5-2 to z = 0 is required to bring these galaxies on to the local scaling relations. The different properties and the different behaviour shown by the two populations with respect to the scaling relations imply different formation and evolution scenarios. The older population of ETGs must have formed at a higher z in a sort of dissipative gas-rich collapse able to produce remnants which at z similar to 2 are old and compact, a scenario which can be fitted qualitatively by some recent hydrodynamic simulations of gas-rich mergers. Given the typical time-scale of merging and the old age of their stellar population, oETGs should exist as they are up to z greater than or similar to 3-3.5. The size evolution they must experience from z similar to 2 to similar to 0 must leave unchanged their mass to not exceed the local number of high-mass (M-* > 5 x 10(11)M(circle dot)) ETGs. Thus, major merging cannot fit this requirement. Satellite merging, close encounters and interactions can help at least qualitatively in solving this problem. The younger population of ETGs can be formed later through subsequent episodes of merging which increased progressively their size and assembled their mass down to z similar to 2. At z < 2, they evolve purely in luminosity since episodes of major merging would bring them far from the local scaling relations.