Yu-Hsuan Hwang (ASIAA); Wei-Hao Wang (ASIAA); Yu-Yen Chang (NCHU); Chen-Fatt Lim (ASIAA); Chian-Chou Chen (ASIAA); and Zhen-Kai Gao (NCU, ASIAA)
The rapid quenching of massive quiescent galaxies (QGs) in the early universe is an important topic in galaxy evolution. We would like to revisit the color-color selection of QGs which is widely applied in QG studies. It is important to examine the robustness of the selection method since the degeneracy of old stars and dust absorption may cause dusty star-forming galaxies (SFGs) contaminating the color-selected QGs. This issue may affect the subsequent analysis and have consequences on our studies of the quenching mechanism. In this study, we selected 17811 QGs using the NUV-r-J diagram out to redshift around 3 from the high-quality multi-wavelength COSMOS2015 catalog. We examined their quiescence using JCMT SCUBA-2 450 and 850μm images from the S2COSMOS and STUDIES Large Programs, which provide the deepest submillimeter view of QGs covering the whole COSMOS field. We used auxiliary data including Spitzer MIPS 24μm, VLA 3GHz, and ALMA AS2COSMOS and A3COSMOS catalogs. The high resolution of auxiliary data and the large sample size of ALMA sources allows us to identify bright submillimeter galaxies among our selected QGs. We examined their spatial correlation and found a clustering signal on a scale of 7”. By stacking the deep SCUBA-2 images, we found a population of 24μm or 3GHz detected (excluding radio AGNs) galaxies distributed out to redshift around 2 which is likely to be contamination of faint dusty SFGs. The results suggest a ∼10% contamination rate in the commonly used NUV-r-J selection, and therefore QG studies that employ similar selections should be made with strong caution. In addition, we found a strong correlation between QGs and radio AGNs in our sample, which may suggest a connection between the quenching mechanism and radio-mode AGN feedback.