Numerical simulations of cosmological structure formation show that the Universes most massive clusters, and the galaxies living in those clusters, assemble rapidly at early times (2.5 < z < 4). While more than twenty proto-clusters have been observe
d at z > 2 based on associations of 5-40 galaxies around rare sources, the observational evidence for rapid cluster formation is weak. Here we report observations of an asymmetric, filamentary structure at z = 2.47 containing seven starbursting, submillimeter-luminous galaxies and five additional AGN within a comoving volume of 15000 Mpc$^{3}$. As the expected lifetime of both the luminous AGN and starburst phase of a galaxy is ~100 Myr, we conclude that these sources were likely triggered in rapid succession by environmental factors, or, alternatively, the duration of these cosmologically rare phenomena is much longer than prior direct measurements suggest. The stellar mass already built up in the structure is $sim10^{12}M_{odot}$ and we estimate that the cluster mass will exceed that of the Coma supercluster at $z sim 0$. The filamentary structure is in line with hierarchical growth simulations which predict that the peak of cluster activity occurs rapidly at z > 2.
Galaxies rest-frame ultraviolet (UV) properties are often used to directly infer the degree to which dust obscuration affects the measurement of star formation rates. While much recent work has focused on calibrating dust attenuation in galaxies sele
cted at rest-frame ultraviolet wavelengths, locally and at high-$z$, here we investigate attenuation in dusty, star-forming galaxies (DSFGs) selected at far-infrared wavelengths. By combining multiwavelength coverage across 0.15--500,$mu$m in the COSMOS field, in particular making use of {it Herschel} imaging, and a rich dataset on local galaxies, we find a empirical variation in the relationship between rest-frame UV slope ($beta$) and ratio of infrared-to-ultraviolet emission ($L_{rm IR}/L_{rm UV}equiv,IRX$) as a function of infrared luminosity, or total star formation rate, SFR. Both locally and at high-$z$, galaxies above SFR$gt$50,M$_odot$,yr$^{-1}$ deviate from the nominal $IRX-beta$ relation towards bluer colors by a factor proportional to their increasing IR luminosity. We also estimate contamination rates of DSFGs on high-$z$ dropout searches of $ll1$% at $zlt4-10$, providing independent verification that contamination from very dusty foreground galaxies is low in LBG searches. Overall, our results are consistent with the physical interpretation that DSFGs, e.g. galaxies with $>50$,M$_odot$,yr$^{-1}$, are dominated at all epochs by short-lived, extreme burst events, producing many young O and B stars that are primarily, yet not entirely, enshrouded in thick dust cocoons. The blue rest-frame UV slopes of DSFGs are inconsistent with the suggestion that most DSFGs at $zsim2$ exhibit steady-state star formation in secular disks.
We present Keck spectroscopic observations and redshifts for a sample of 767 Herschel-SPIRE selected galaxies (HSGs) at 250, 350, and 500um, taken with the Keck I Low Resolution Imaging Spectrometer (LRIS) and the Keck II DEep Imaging Multi-Object Sp
ectrograph (DEIMOS). The redshift distribution of these SPIRE sources from the Herschel Multitiered Extragalactic Survey (HerMES) peaks at z=0.85, with 731 sources at z<2 and a tail of sources out to z~5. We measure more significant disagreement between photometric and spectroscopic redshifts (<delta_z>/(1+z)>=0.29) than is seen in non-infrared selected samples, likely due to enhanced star formation rates and dust obscuration in infrared-selected galaxies. We estimate that the vast majority (72-83%) of z<2 Herschel-selected galaxies would drop out of traditional submillimeter surveys at 0.85-1mm. We estimate the luminosity function and implied star-formation rate density contribution of HSGs at z<1.6 and find overall agreement with work based on 24um extrapolations of the LIRG, ULIRG and total infrared contributions. This work significantly increased the number of spectroscopically confirmed infrared-luminous galaxies at z>>0 and demonstrates the growing importance of dusty starbursts for galaxy evolution studies and the build-up of stellar mass throughout cosmic time. [abridged]
We present near-infrared spectroscopic observations from VLT ISAAC of thirteen 250mu m-luminous galaxies in the CDF-S, seven of which have confirmed redshifts which average to <z > = 2.0 pm 0.4. Another two sources of the 13 have tentative z > 1 iden
tifications. Eight of the nine redshifts were identified with H{alpha} detection in H- and K-bands, three of which are confirmed redshifts from previous spectroscopic surveys. We use their near-IR spectra to measure H{alpha} line widths and luminosities, which average to 415 pm 20 km/s and 3 times 10^35 W (implying SFR(H{alpha})~200 M_odot /yr), both similar to the H{alpha} properties of SMGs. Just like SMGs, 250 mu m-luminous galaxies have large H{alpha} to far-infrared (FIR) extinction factors such that the H{alpha} SFRs underestimate the FIR SFRs by ~8-80 times. Far-infrared photometric points from observed 24mu m through 870mu m are used to constrain the spectral energy distributions (SEDs) even though uncertainty caused by FIR confusion in the BLAST bands is significant. The population has a mean dust temperature of Td = 52 pm 6 K, emissivity {beta} = 1.73 pm 0.13, and FIR luminosity LFIR = 3 times 10^13 L_odot. Although selection at 250mu m allows for the detection of much hotter dust dominated HyLIRGs than SMG selection (at 850mu m), we do not find any >60 K hot-dust HyLIRGs. We have shown that near-infrared spectroscopy combined with good photometric redshifts is an efficient way to spectroscopically identify and characterise these rare, extreme systems, hundreds of which are being discovered by the newest generation of IR observatories including the Herschel Space Observatory.
Using the IRAM Plateau-de-Bure Interferometer (PdBI) we have searched for the upper fine structure line of neutral carbon (CI(2->1), nu_rest = 809 GHz) and 12CO(J=7->6) (nu_rest=806GHz) towards the submillimetre galaxies (SMGs) GN20 (SMM J123711.9+62
2212, z = 4.055) and GN20.2 (SMM J123708.8+622202, z = 4.051). The far-infrared (FIR) continuum is detected at 8sigma significance in GN20, with a flux density of S_1.8mm = 1.9+-0.2mJy, while no continuum is detected in GN20.2. Both sources are statistically undetected in both CI and CO(7-6) lines; we derive line luminosity limits for both CI and CO of L<~2x10^10K km s^-1 pc^2. Assuming carbon excitation temperatures of T_ex = 30K (the galaxies measured dust temperatures), we infer CI mass limits of M_CI < 5.4x10^6 Msun (GN20) and M_CI < 6.8x10^6 Msun (GN20.2). The derived CI abundance limits are <1.8x10^-5 for GN20 and <3.8x10^-5 for GN20.2 implying that the systems have Milky Way level neutral carbon enrichment (X[CI]/X[H2]) or lower, similar to high-redshift carbon-detected systems (at 5x10^-5) but about 50 times less than the neutral carbon enrichment of local starburst galaxies. Observations of GN20 and GN20.2 in high-resolution MERLIN+VLA radio maps of GOODS-N are used to further constrain the sizes and locations of active regions. We conclude that the physical gas properties of young rapidly evolving systems like GN20 and GN20.2 are likely significantly different than starburst/ULIRG environments in the local Universe yet similar to z~2 SMGs. Unless gravitationally amplified examples can be found, observations of galaxies like GN20 will require the order of magnitude increase in sensitivity of the Atacama Large Millimetre Array (ALMA) to constrain their CI and high-J CO content, despite the fact that they are the brightest systems at z~4.
We outline a strategy to select faint (i<24.5) type 1 AGN candidates down to the Seyfert/QSO boundary for spectroscopic targeting in the COSMOS field, picking candidates by their nonstellar colors in broadband ground-based photometry and morphologica
l properties extracted from HST-ACS. AGN optical color selection has not been applied to such faint magnitudes in such a large continuous part of the sky. Hot stars are known to be the dominant contaminant for bright AGN candidate selection at z<2, but we anticipate the highest color contamination at all redshifts to be from faint starburst and compact galaxies. Morphological selection via the Gini Coefficient separates most potential AGN from these faint blue galaxies. Recent models of the quasar luminosity function are used to estimate quasar surface densities, and studies of stellar populations in the COSMOS field infer stellar contamination. We use 292 spectroscopically confirmed type 1 AGN and quasar templates to predict AGN colors with redshift, and contrast those predictions with the colors of known contaminating populations. The motivation of this study and subsequent spectroscopic follow-up is to populate and refine the faint end of the QLF where the population of type 1 AGN is presently not well known. The anticipated AGN observations will add to the ~300 already known AGN in the COSMOS field, making COSMOS a densely packed field of quasars to be used to understand supermassive black holes and probe the structure of the intergalactic medium in the intervening volume.
