We investigate the nature and the physical properties of z, Y and J-dropout galaxies selected behind the lensing cluster A2667. This field is part of our project aimed at identifying z~7-10 candidates accessible to spectroscopic studies, based on deep photometry with ESO/VLT HAWK-I and FORS2 (zYJH and Ks-band images, AB(3 sigma)~26-27) on a sample of lensing clusters extracted from our multi-wavelength combined surveys with SPITZER, HST, and Herschel. In this paper we focus on the complete Y and J-dropout sample, as well as the bright z-dropouts fulfilling the selection criteria by Capak et al. (2011). 10 candidates are selected within the common field of ~33 arcmin2 (effective area once corrected for contamination and lensing dilution). All of them are detected in H and Ks bands in addition to J and/or IRAC 3.6/4.5, with H(AB)~23.4 to 25.2, and have modest magnification factors. Although best-fit photometric redshifts place all these candidates at high-z, the contamination by low-z interlopers is estimated at 50-75% level based on previous studies, and the comparison with the blank-field WIRCAM Ultra-Deep Survey (WUDS). The same result is obtained when photometric redshifts include a luminosity prior, allowing us to remove half of the original sample as likely z~1.7-3 interlopers with young stellar pulations and strong extinction. Two additional sources among the remaining sample could be identified at low-z based on a detection at 24 microns and on the HST z_850 band. These low-z interlopers are not well described by current templates given the large break, and cannot be easily identified based solely on optical and near-IR photometry. Given the estimated dust extinction and high SFRs, some of them could be also detected in the IR or sub-mm bands. After correction for likely contaminants, the observed counts at z>7.5 seem to be in agreement with an evolving LF. (abridged)

The identification and study of the first galaxies remains one of the most exciting topics in observational cosmology. The determination of the best possible observing strategies is a very important choice in order to build up a representative sample of spectroscopically confirmed sources at high-z (z>7), beyond the limits of present-day observations. This paper is intended to precisely adress the relative efficiency of lensing and blank fields in the identification and study of galaxies at 6<z<12. The detection efficiency and field-to-field variance are estimated from direct simulations of both blank and lensing fields observations. The presence of a strong-lensing cluster along the line of sight has a dramatic effect on the number of observed sources, with a positive magnification bias in typical ground-based ``shallow'' surveys (AB<~25.5). The positive magnification bias increases with the redshift of sources and decreases with both depth of the survey and the size of the surveyed area. The maximum efficiency is reached for lensing clusters at z~0.1-0.3. Observing blank fields in shallow surveys is particularly inefficient as compared to lensing fields if the UV LF for LBGs is strongly evolving at z>~7. Also in this case, the number of z>8 sources expected at the typical depth of JWST (AB~28-29) is much higher in lensing than in blank fields (e.g. a factor of ~10 for AB<~28). Blank field surveys with a large field of view are needed to prove the bright end of the LF at z>6-7, whereas lensing clusters are particularly useful for exploring the mid to faint end of the LF.