The Herschel Multi-tiered Extragalactic Survey (HerMES) has identified large numbers of dusty star-forming galaxies (DSFGs) over a wide range in redshift. A detailed understanding of these DSFGs is hampered by the limited spatial resolution of Hersch
el. We present 870um 0.45 resolution imaging from the Atacama Large Millimeter/submillimeter Array (ALMA) of 29 HerMES DSFGs with far-infrared (FIR) flux densities in between the brightest of sources found by Herschel and fainter DSFGs found in ground-based sub-millimeter (sub-mm) surveys. We identify 62 sources down to the 5-sigma point-source sensitivity limit in our ALMA sample (sigma~0.2mJy), of which 6 are strongly lensed (showing multiple images) and 36 experience significant amplification (mu>1.1). To characterize the properties of the ALMA sources, we introduce and make use of uvmcmcfit, a publicly available Markov chain Monte Carlo analysis tool for interferometric observations of lensed galaxies. Our lens models tentatively favor intrinsic number counts for DSFGs with a steep fall off above 8mJy at 880um. Nearly 70% of the Herschel sources comprise multiple ALMA counterparts, consistent with previous research indicating that the multiplicity rate is high in bright sub-mm sources. Our ALMA sources are located significantly closer to each other than expected based on results from theoretical models as well as fainter DSFGs identified in the LABOCA ECDFS Submillimeter Survey. The high multiplicity rate and low projected separations argue in favor of interactions and mergers driving the prodigious emission from the brightest DSFGs as well as the sharp downturn above S_880=8mJy.
Strong gravitational lenses are now being routinely discovered in wide-field surveys at (sub)millimeter wavelengths. We present Submillimeter Array (SMA) high-spatial resolution imaging and Gemini-South and Multiple Mirror Telescope optical spectrosc
opy of strong lens candidates discovered in the two widest extragalactic surveys conducted by the Herschel Space Observatory: the Herschel-Astrophysical Terahertz Large Area Survey (H-ATLAS) and the Herschel Multi-tiered Extragalactic Survey (HerMES). From a sample of 30 Herschel sources with S_500>100 mJy, 21 are strongly lensed (multiply imaged), 4 are moderately lensed (singly imaged), and the remainder require additional data to determine their lensing status. We apply a visibility-plane lens modeling technique to the SMA data to recover information about the masses of the lenses as well as the intrinsic (i.e., unlensed) sizes (r_half) and far-infrared luminosities (L_FIR) of the lensed submillimeter galaxies (SMGs). The sample of lenses comprises primarily isolated massive galaxies, but includes some groups and clusters as well. Several of the lenses are located at z_lens>0.7, a redshift regime that is inaccessible to lens searches based on Sloan Digital Sky Survey spectroscopy. The lensed SMGs are amplified by factors that are significantly below statistical model predictions given the 500um flux densities of our sample. We speculate that this may reflect a deficiency in our understanding of the intrinsic sizes and luminosities of the brightest SMGs. The lensed SMGs span nearly one decade in L_FIR (median L_FIR=7.9x10^12 L_sun) and two decades in FIR luminosity surface density (median Sigma_FIR=6.0x10^11 L_sun kpc^-2). The strong lenses in this sample and others identified via (sub-)mm surveys will provide a wealth of information regarding the astrophysics of galaxy formation and evolution over a wide range in redshift.
We present high-spatial resolution imaging obtained with the Submillimeter Array (SMA) at 880um and the Keck Adaptive Optics (AO) system at Ks-band of a gravitationally lensed sub-millimeter galaxy (SMG) at z=4.243 discovered in the Herschel-Astrophy
sical Terahertz Large Area Survey. The SMA data (angular resolution ~0.6) resolve the dust emission into multiple lensed images, while the Keck AO Ks-band data (angular resolution ~0.1) resolve the lens into a pair of galaxies separated by 0.3. We present an optical spectrum of the foreground lens obtained with the Gemini-South telescope that provides a lens redshift of z_lens = 0.595 +/- 0.005. We develop and apply a new lens modeling technique in the visibility plane that shows that the SMG is magnified by a factor of mu = 4.1 +/- 0.2 and has an intrinsic infrared (IR) luminosity of L_IR = (2.1 +/- 0.2) x 10^13 Lsun. We measure a half-light radius of the background source of r_s = 4.4 +/- 0.5 kpc which implies an IR luminosity surface density of Sigma_IR = (3.4 +/- 0.9) x 10^11 Lsun kpc^-2, a value that is typical of z > 2 SMGs but significantly lower than IR luminous galaxies at z~0. The two lens galaxies are compact (r_lens ~ 0.9 kpc) early-types with Einstein radii of theta_E1 = 0.57 +/- 0.01 and theta_E2 = 0.40 +/- 0.01 that imply masses of M_lens1 = (7.4 +/- 0.5) x 10^10 Msun and M_lens2 = (3.7 +/- 0.3) x 10^10 Msun. The two lensing galaxies are likely about to undergo a dissipationless merger, and the mass and size of the resultant system should be similar to other early-type galaxies at z~0.6. This work highlights the importance of high spatial resolution imaging in developing models of strongly lensed galaxies discovered by Herschel.
The Spitzer Space Telescope has identified a population of ultra-luminous infrared galaxies (ULIRGs) at z ~ 2 that may play an important role in the evolution of massive galaxies. We measure the stellar masses of two populations of Spitzer-selected U
LIRGs, both of which have extremely red R-[24] colors (dust-obscured galaxies, or DOGs) and compare our results with sub-millimeter selected galaxies (SMGs). One set of 39 DOGs has a local maximum in their mid-IR spectral energy distribution (SED) at rest-frame 1.6um associated with stellar emission (bump DOGs), while the other set of 51 DOGs has a power-law dominated mid-IR SED with spectral features typical of obscured AGN (power-law DOGs). We use stellar population synthesis models applied self-consistently to broad-band photometry in the rest-frame ultra-violet, optical, and near-infrared of each of these populations and test a variety of stellar population synthesis codes, star-formation histories (SFHs), and initial mass functions (IMFs). Assuming a simple stellar population SFH and a Chabrier IMF, we find that the median and inner quartile stellar masses of SMGs, bump DOGs and power-law DOGs are given by log(M_*/M_sun) = 10.42_-0.36^+0.42, 10.62_-0.32^+0.36, and 10.71_-0.34^+0.40, respectively. Implementing more complicated SFHs with multiple age components increases these mass estimates by up to 0.5 dex. Our stellar mass estimates are consistent with physical mechanisms for the origin of z~2 ULIRGs that result in high star-formation rates for a given stellar mass. Such mechanisms are usually driven by a major merger of two gas-rich systems, rather than smooth accretion of gas and small satellites.
We present Hubble Space Telescope (HST) imaging of 22 ultra-luminous infrared galaxies (ULIRGs) at z~2 with extremely red R-[24] colors (called dust-obscured galaxies, or DOGs) which have a local maximum in their spectral energy distribution (SED) at
rest-frame 1.6um associated with stellar emission. These sources, which we call bump DOGs, have star-formation rates of 400-4000 Msun/yr and have redshifts derived from mid-IR spectra which show strong polycyclic aromatic hydrocarbon emission --- a sign of vigorous on-going star-formation. Using a uniform morphological analysis, we look for quantifiable differences between bump DOGs, power-law DOGs (Spitzer-selected ULIRGs with mid-IR SEDs dominated by a power-law and spectral features that are more typical of obscured active galactic nuclei than starbursts), sub-millimeter selected galaxies (SMGs), and other less-reddened ULIRGs from the Spitzer extragalactic First Look Survey (XFLS). Bump DOGs are larger than power-law DOGs (median Petrosian radius of 8.4 +/- 2.7 kpc vs. 5.5 +/- 2.3 kpc) and exhibit more diffuse and irregular morphologies (median M_20 of -1.08 +/- 0.05 vs. -1.48 +/- 0.05). These trends are qualitatively consistent with expectations from simulations of major mergers in which merging systems during the peak star-formation rate period evolve from M_20 = -1.0 to M_20 = -1.7. Less obscured ULIRGs (i.e., non-DOGs) tend to have more regular, centrally peaked, single-object morphologies rather than diffuse and irregular morphologies. This distinction in morphologies may imply that less obscured ULIRGs sample the merger near the end of the peak star-formation rate period. Alternatively, it may indicate that the intense star-formation in these less-obscured ULIRGs is not the result of a recent major merger.
We present SHARC-II 350um imaging of twelve 24um-bright (F_24um > 0.8 mJy) Dust-Obscured Galaxies (DOGs) and CARMA 1mm imaging of a subset of 2 DOGs, all selected from the Bootes field of the NOAO Deep Wide-Field Survey. Detections of 4 DOGs at 350um
imply IR luminosities which are consistent within a factor of 2 of expectations based on a warm dust spectral energy distribution (SED) scaled to the observed 24um flux density. The 350um upper limits for the 8 non-detected DOGs are consistent with both Mrk231 and M82 (warm dust SEDs), but exclude cold dust (Arp220) SEDs. The two DOGs targeted at 1mm were not detected in our CARMA observations, placing strong constraints on the dust temperature: T_dust > 35-60 K. Assuming these dust properties apply to the entire sample, we find dust masses of ~3x10^8 M_sun. In comparison to other dusty z ~ 2 galaxy populations such as sub-millimeter galaxies (SMGs) and other Spitzer-selected high-redshift sources, this sample of DOGs has higher IR luminosities (2x10^13 L_sun vs. 6x10^12 L_sun for the other galaxy populations), warmer dust temperatures (>35-60 K vs. ~30 K), and lower inferred dust masses (3x10^8 M_sun vs. 3x10^9 M_sun). Herschel and SCUBA-2 surveys should be able to detect hundreds of these power-law dominated DOGs. We use HST and Spitzer/IRAC data to estimate stellar masses of these sources and find that the stellar to gas mass ratio may be higher in our 24um-bright sample of DOGs than in SMGs and other Spitzer-selected sources. Although larger sample sizes are needed to provide a definitive conclusion, the data are consistent with an evolutionary trend in which the formation of massive galaxies at z~2 involves a sub-millimeter bright, cold-dust and star-formation dominated phase followed by a 24um-bright, warm-dust and AGN-dominated phase.
We present high spatial resolution optical and near-infrared imaging obtained using the ACS, WFPC2 and NICMOS cameras aboard the Hubble Space Telescope of 31 24um--bright z~2 Dust Obscured Galaxies (DOGs) identified in the Bootes Field of the NOAO De
ep Wide-Field Survey. Although this subset of DOGs have mid-IR spectral energy distributions dominated by a power-law component suggestive of an AGN, all but one of the galaxies are spatially extended and not dominated by an unresolved component at rest-frame UV or optical wavelengths. The observed V-H and I-H colors of the extended components are 0.2-3 magnitudes redder than normal star-forming galaxies. All but 1 have axial ratios >0.3, making it unlikely that DOGs are composed of an edge-on star-forming disk. We model the spatially extended component of the surface brightness distributions of the DOGs with a Sersic profile and find effective radii of 1-6 kpc. This sample of DOGs is smaller than most sub-millimeter galaxies (SMGs), but larger than quiescent high-redshift galaxies. Non-parametric measures (Gini and M20) of DOG morphologies suggest that these galaxies are more dynamically relaxed than local ULIRGs. We estimate lower limits to the stellar masses of DOGs based on the rest-frame optical photometry and find that these range from ~10^(9-11) M_sun. If major mergers are the progenitors of DOGs, then these observations suggest that DOGs may represent a post-merger evolutionary stage.
We present observations made with the 10m Heinrich Hertz Sub-Millimeter Telescope of HCN (3-2) emission from a sample of 30 nearby galaxies ranging in infrared luminosity from 10^10 - 10^12.5 L_sun and HCN (3-2) luminosity from 10^6 - 10^9 K km s^-1
pc^2. We examine the correlation between the infrared luminosity and HCN (3-2) luminosity and find that the best fit linear regression has a slope (in log-log space) of 0.74+/-0.12. Including recently published data from Gracia-Carpio et al. tightens the constraints on the best-fit slope to 0.79+/-0.09. This slope below unity suggests that the HCN (3-2) molecular line luminosity is not linearly tracing the amount of dense gas. Our results are consistent with predictions from recent theoretical models that find slopes below unity when the line luminosity depends upon the average gas density with a power-law index greater than a Kennicutt-Schmidt index of 1.5.
