Seth H. Cohen

Galaxy formation in the reionization epoch as hinted by Wide Field Camera 3 observations of the Hubble Ultra Deep Field

We present a large sample of candidate galaxies at z ≈ 7–10, selected in the Hubble Ultra Deep Field using the new observations of the Wide Field Camera 3 that was recently installed on the Hubble Space Telescope. Our sample is composed of 20 z850-dropouts (four new discoveries), 15 Y105-dropouts (nine new discoveries) and 20 J125-dropouts (all new discoveries). The surface densities of the z850-dropouts are close to what was predicted by earlier studies, however, those of the Y105- and J125-dropouts are quite unexpected. While no Y105- or J125-dropouts have been found at AB ≤ 28.0 mag, their surface densities seem to increase sharply at fainter levels. While some of these candidates seem to be close to foreground galaxies and thus could possibly be gravitationally lensed, the overall surface densities after excluding such cases are still much higher than what would be expected if the luminosity function does not evolve from z ~ 7 to 10. Motivated by such steep increases, we tentatively propose a set of Schechter function parameters to describe the luminosity functions at z ≈ 8 and 10. As compared to their counterpart at z ≈ 7, here L* decreases by a factor of ~ 6.5 and * increases by a factor of 17–90. Although such parameters are not yet demanded by the existing observations, they are allowed and seem to agree with the data better than other alternatives. If these luminosity functions are still valid beyond our current detection limit, this would imply a sudden emergence of a large number of low-luminosity galaxies when looking back in time to z ≈ 10, which, while seemingly exotic, would naturally fit in the picture of the cosmic hydrogen reionization. These early galaxies could easily account for the ionizing photon budget required by the reionization, and they would imply that the global star formation rate density might start from a very high value at z ≈ 10, rapidly reach the minimum at z ≈ 7, and start to rise again towards z ≈ 6. In this scenario, the majority of the stellar mass that the universe assembled through the reionization epoch seems still undetected by current observations at z ≈ 6.

UV-DROPOUT GALAXIES IN THE GOODS-SOUTH FIELD FROM WFC3 EARLY RELEASE SCIENCE OBSERVATIONS

We combine new high sensitivity ultraviolet (UV) imaging from the Wide Field Camera 3 (WFC3) on the Hubble Space Telescope (HST) with existing deep HST/Advanced Camera for Surveys optical images from the Great Observatories Origins Deep Survey (GOODS) program to identify UV-dropouts, which are Lyman break galaxy (LBG) candidates at z 1-3. These new HST/WFC3 observations were taken over 50 arcmin2 in the GOODS-South field as a part of the Early Release Science program. The uniqueness of these new UV data is that they are observed in three UV/optical (WFC3 UVIS) channel filters (F225W, F275W, and F336W), which allows us to identify three different sets of UV-dropout samples. We apply Lyman break dropout selection criteria to identify F225W-, F275W-, and F336W-dropouts, which are z 1.7, 2.1, and 2.7 LBG candidates, respectively. We use multi-wavelength imaging combined with available spectroscopic and photometric redshifts to carefully access the validity of our UV-dropout candidates. Our results are as follows: (1) these WFC3 UVIS filters are very reliable in selecting LBGs with z 2.0, which helps to reduce the gap between the well-studied z 3 and z ~ 0 regimes; (2) the combined number counts with average redshift z 2.2 agree very well with the observed change in the surface densities as a function of redshift when compared with the higher redshift LBG samples; and (3) the best-fit Schechter function parameters from the rest-frame UV luminosity functions at three different redshifts fit very well with the evolutionary trend of the characteristic absolute magnitude, M*, and the faint-end slope, ?, as a function of redshift. This is the first study to illustrate the usefulness of the WFC3 UVIS channel observations to select z 3 LBGs. The addition of the new WFC3 on the HST has made it possible to uniformly select LBGs from z 1 to z 9 and significantly enhance our understanding of these galaxies using HST sensitivity and resolution.

Galaxy mergers at z ≳ 1 in the HUDF: Evidence for a peak in the major merger rate of massive galaxies

We present a measurement of the galaxy merger fraction and number density from observations in the Hubble Ultra Deep Field (HUDF) for -->0.5 ? z ? 2.5. We fit the combination of broadband data and slitless spectroscopy of 1308 galaxies with stellar population synthesis models to select merging systems based on a stellar mass of ?1010 -->M?. When correcting for mass incompleteness, the major merger fraction is not simply proportional to -->(1 + z)m but appears to peak at -->zfrac 1.3 ? 0.4. From this merger fraction, we infer that ~42% of massive galaxies have undergone a major merger since -->z 1. We show that the major merger number density peaks at -->zdens 1.2, which marks the epoch where major merging of massive galaxies is most prevalent. This critical redshift is comparable to the peak of the cosmic star formation rate density and occurs roughly 2.6 Gyr earlier in cosmic time than the peak in the number density of the X-ray-selected active galactic nuclei. These observations support an indirect evolutionary link between merging, starburst, and active galaxies.

Automated Galaxy Morphology: A Fourier Approach

We use automated surface photometry and pattern classification techniques to morphologically classify galaxies. The two-dimensional light distribution of a galaxy is reconstructed using Fourier series fits to azimuthal profiles computed in concentric elliptical annuli centered on the galaxy. Both the phase and amplitude of each Fourier component have been studied as a function of radial bin number for a large collection of galaxy images using principal-component analysis. We find that up to 90% of the variance in many of these Fourier profiles may be characterized in as few as three principal components and that their use substantially reduces the dimensionality of the classification problem. We use supervised learning methods in the form of artificial neural networks to train galaxy classifiers that detect morphological bars at the 85%-90% confidence level and can identify the Hubble type with a 1 σ scatter of 1.5 steps on the 16 step stage axis of the revised Hubble system. Finally, we systematically characterize the adverse effects of decreasing resolution and signal-to-noise ratio on the quality of morphological information predicted by these classifiers.

The insignificance of major mergers in driving star formation at z ≃ 2

We study the significance of major-merger-driven star formation in the early Universe, by quantifying the contribution of this process to the total star formation budget in 80 massive (M∗ > 10 10 M⊙) galaxies at z ≃ 2. Employing visually-classified morphologies from rest-frame V -band HST imaging, we find that 55 ±14 % of the star formation budget is hosted by non-interacting late-types, with 27 ±8 % in major mergers and 18 ±6 % in spheroids. Given that a system undergoing a major merger continues to experience star formation driven by other processes at this epoch (e.g. cold accretion, minor mergers), ∼27% is an upper limit to the major-merger contribution to star formation activity at this epoch. The ratio of the average specific star formation rate in major mergers to that in the non-interacting late-types is ∼2.2:1, suggesting that the enhancement of star formation due to major merging is typically modest, and that just under half the star formation in systems experiencing major mergers is unrelated to the merger itself. Taking this into account, we estimate that the actual majormerger contribution to the star formation budget may be as low as ∼15%. While our study does not preclude a major-merger-dominated era in the very early Universe, if the major-merger contribution to star formation does not evolve strongly into larger look-back times, then this process has a relatively insignificant role in driving stellar mass assembly over cosmic time.

NICMOS Imaging of the Dusty Microjansky Radio Source VLA J123642+621331 at z = 4.424*

We present the discovery of a radio galaxy at a likely redshift of z=4.424 in one of the flanking fields of the Hubble Deep Field (HDF). Radio observations with the Very Large Array and MERLIN centered on the HDF yielded a complete sample of microjansky radio sources, of which about 20% have no optical counterpart to I</=25 mag. In this Letter, we address the possible nature of one of these sources through deep Hubble Space Telescope near-infrared camera and multiobject spectrometer (NICMOS) images in the F110W (J110) and F160W (H160) filters. VLA J123642+621331 has a single emission line at 6595 Å, which we identify with Lyalpha at z=4.424. We argue that this faint (H160=23.9 mag), compact (re approximately 0&farcs;2), red (I814-K=2.0) object is most likely a dusty, star-forming galaxy with an embedded active nucleus.

PHYSICAL PROPERTIES OF SPECTROSCOPICALLY CONFIRMED GALAXIES AT z ⩾ 6. I. BASIC CHARACTERISTICS OF THE REST-FRAME UV CONTINUUM AND Lyα EMISSION*

We present deep Hubble Space Telescope near-IR and Spitzer mid-IR observations of a large sample of spectroscopically confirmed galaxies at z ≥ 6. The sample consists of 51 Lyα emitters (LAEs) at z 5.7, 6.5, and 7.0, and 16 Lyman break galaxies (LBGs) at 5.9 ≤ z ≤ 6.5. The near-IR images were mostly obtained with WFC3 in the F125W and F160W bands, and the mid-IR images were obtained with IRAC in the 3.6 μm and 4.5 μm bands. Our galaxies also have deep optical imaging data from Subaru Suprime-Cam. We utilize the multi-band data and secure redshifts to derive their rest-frame UV properties. These galaxies have steep UV-continuum slopes roughly between β –1.5 and –3.5, with an average value of β –2.3, slightly steeper than the slopes of LBGs in previous studies. The slope shows little dependence on UV-continuum luminosity except for a few of the brightest galaxies. We find a statistically significant excess of galaxies with slopes around β –3, suggesting the existence of very young stellar populations with extremely low metallicity and dust content. Our galaxies have moderately strong rest-frame Lyα equivalent width (EW) in a range of ~10 to ~200 A. The star formation rates are also moderate, from a few to a few tens of solar masses per year. The LAEs and LBGs in this sample share many common properties, implying that LAEs represent a subset of LBGs with strong Lyα emission. Finally, the comparison of the UV luminosity functions between LAEs and LBGs suggests that there exists a substantial population of faint galaxies with weak Lyα emission (EW < 20 A) that could be the dominant contribution to the total ionizing flux at z ≥ 6.

The Galaxy Luminosity Function at z ≃ 1 in the HUDF: Probing the Dwarf Population*

We present a catalog of spectrophotometric redshifts for 1308 galaxies from the Grism ACS Program for Extragalactic Science (GRAPES) observations with the Hubble Space Telescope. These low-resolution spectra between 6000 A and 9500 A are supplemented with U, J, H, and Ks data from various facilities, resulting in redshifts computed with ~40 spectral bins per galaxy. For 75 galaxies in the range 0.5 < z < 1.5 with spectroscopic redshifts, the standard deviation in the fractional error in 1 + z is 0.037. With this catalog, we compute the B-band luminosity function in this redshift range from 72 galaxies. Due to the depth of the GRAPES survey, we are able to accurately constrain the faint-end slope by going to MB -18 mag at z = 1.0 ± 0.2, nearly 2 mag fainter than previous studies. The faint-end slope is α = -1.32 ± 0.07. When compared to numerous published values at various redshifts, we find strong evidence for a steepening of the faint-end slope with redshift, which is expected in the hierarchical formation scenario of galaxies.

Dust extinction and metallicities of star-forming lyα emitting galaxies at low redshift

We present the results of an optical spectroscopic study of 12 GALEX-discovered star-forming Lyα emitting galaxies (LAEs) at z ~ 0.3. We measure the emission-line fluxes from these galaxies by fitting their observed spectra to stellar population models in order to correct for underlying stellar absorption. We revisit earlier stellar population model fitting results, finding that excluding now-known active galactic nuclei lowers the typical stellar population age and stellar mass of this sample to ~300 Myr and ~4 × 109 ☉, respectively. We calculate their dust extinction using the Balmer decrement, and find a typical visual attenuation of A V ~ 0.3 mag, similar to that seen in some high-redshift LAEs. Comparing the ratios of Lyα/Hα and the Lyα equivalent widths to the measured dust extinction, we find that the interstellar media (ISMs) in these objects appear to be neither enhancing nor seriously attenuating the Lyα equivalent widths, as would be the case in a quasi-clumpy ISM. Lastly, we perform a detailed analysis of the gas-phase metallicities of these galaxies, and we find that most galaxies in our sample have Z 0.4 Z ☉. We find that at a fixed stellar mass, these low-redshift LAE analogs are offset by ~0.3-0.6 dex lower metallicity from the general galaxy population at similar redshifts based on the local mass-metallicity relationship. This implies that galaxies with Lyα in emission may be systematically more metal-poor than star-forming galaxies at the same stellar mass and redshift, similar to preliminary results at z ~ 2.

Newborn spheroids at high redshift: when and how did the dominant, old stars in today's massive galaxies form?

We study ∼330 massive (M* > 109.5 M_⊙), newborn spheroidal galaxies (SGs) around the epoch of peak star formation (1 10^(10.5) M_⊙, and an age trend becomes evident in this mass regime: SGs with M* > 10^(11.5) M_⊙ are ∼2 Gyr older than their counterparts with M* < 10^(10.5) M_⊙. Nevertheless, a smooth downsizing trend with galaxy mass is not observed, and the large scatter in starburst ages indicates that SGs are not a particularly coeval population. Around half of the blue SGs appear not to drive their star formation via major mergers, and those that have experienced a recent major merger show only modest enhancements (∼40 per cent) in their specific star formation rates. Our empirical study indicates that processes other than major mergers (e.g. violent disc instability driven by cold streams and/or minor mergers) likely play a dominant role in building SGs, and creating a significant fraction of the old stellar populations that dominate todays Universe.