Passively evolving early-type galaxies at 1.4 ≲ z ≲ 2.5 in the Hubble Ultra Deep Field

We report on a complete sample of seven luminous early-type galaxies in the Hubble Ultra Deep Field (UDF) with spectroscopic redshifts between 1.39 and 2.47, and to K_AB < 23. Using the BzK selection criterion, we have preselected a set of objects over the UDF, which fulfill the photometric conditions for being passively evolving galaxies at z > 1.4. Low-resolution spectra of these objects have been extracted from the Hubble Space Telescope (HST) ACS grism data taken over the UDF by the Grism ACS Program for Extragalactic Science (GRAPES) project. Redshifts for the seven galaxies have been identified based on the UV feature at rest frame 2640 < λ < 2850 Å. This feature is mainly due to a combination of Fe II, Mg I, and Mg II absorptions, which are characteristic of stellar populations dominated by stars older than ∼0.5 Gyr. The redshift identification and the passively evolving nature of these galaxies is further supported by the photometric redshifts and by the overall spectral energy distribution (SED), with the ultradeep HST ACS NICMOS imaging revealing compact morphologies typical of elliptical/early-type galaxies. From the SED we derive stellar masses of ≳10¹¹ M_⊙ and ages of ∼1 Gyr. Their space density at 〈z〉 = 1.7 appears to be roughly a factor of 2-3 smaller than that of their local counterparts, further supporting the notion that such massive and old galaxies are already ubiquitous at early cosmic times. Much smaller effective radii are derived for some of the objects, compared to local massive ellipticals, which may be due to morphological K-corrections, evolution, or the presence of a central pointlike source. Nuclear activity is indeed present in a subset of the galaxies, as revealed by the fact that they are hard X-ray sources, which suggests that active galactic nucleus (AGN) activity may have played a role in discontinuing star formation.