Search

International audience; We analyze the physical conditions in the interstellar gas of 11 actively star-forming galaxies at z~2, based on integral-field spectroscopy from the ESO-VLT and Hubble Space Telescope/Near Infrared Camera and Multi-Object Spectrometer imaging. We concentrate on the high Halpha surface brightnesses, large line widths, line ratios, and the clumpy nature of these galaxies. We show that photoionization calculations and emission line diagnostics imply gas pressures and densities that are similar to the most intense nearby star-forming regions at z = 0 but over much larger scales (10-20 kpc). A relationship between surface brightness and velocity dispersion can be explained through simple energy injection arguments and a scaling set by nearby galaxies with no free parameters. The high velocity dispersions are a natural consequence of intense star formation thus regions of high velocity dispersion are not evidence for mass concentrations such as bulges or rings. External mechanisms such as cosmological gas accretion generally do not have enough energy to sustain the high velocity dispersions. In some cases, the high pressures and low gas metallicites may make it difficult to robustly distinguish between active galactic nucleus ionization cones and star formation, as we show for BzK-15504 at z = 2.38. We construct a picture where the early stages of galaxy evolution are driven by self-gravity which powers strong turbulence until the velocity dispersion is high. Then massive, dense, gas-rich clumps collapse, triggering star formation with high efficiencies and intensities as observed. At this stage, the intense star formation is likely self-regulated by the mechanical energy output of massive stars.

[Abridged]We compare both the Milky Way and M31 galaxies to local external disk galaxies within the same mass range, using their relative locations in the planes formed by V_flat versus M_K, j_disk, and the average Fe abundance of stars in the galaxy outskirts. We find, for all relationships, that the MW is systematically offset by ~ 1 sigma, showing a significant deficiency in stellar mass, in angular momentum, in disk radius and [Fe/H] in the stars in its outskirts at a given V_flat. On the basis of their location in the M_K, V_flat, and R_d volume, the fraction of spirals like the MW is 7+/-1%, while M31 appears to be a "typical'' spiral. Our Galaxy appears to have escaped any significant merger over the last ~10 Gyrs which may explain why it is deficient by a factor 2 to 3 in stellar mass, angular momentum and outskirts metallicity and then, unrepresentative of the typical spiral. As with M31, most local spirals show evidence for a history shaped mainly by relatively recent merging. We conclude that the standard scenario of secular evolution is generally unable to reproduce the properties of most (if not all) spiral galaxies. However, the so-called "spiral rebuilding'' scenario proposed by Hammer et al. 2005 is consistent with the properties of both distant galaxies and of their descendants - the local spirals., Comment: 14 pages, 6 figures, to appear in ApJ

We present rest frame mid-infrared spectroscopy of a sample of 13 submillimeter galaxies, obtained using the Infrared Spectrograph (IRS) on board the Spitzer Space Telescope. The sample includes exclusively bright objects from blank fields and cluster lens assisted surveys that have accurate interferometric positions. We find that the majority of spectra are well fitted by a starburst template or by the superposition of PAH emission features and a weak mid-infrared continuum, the latter a tracer of Active Galactic Nuclei (including Compton-thick ones). We obtain mid-infrared spectroscopic redshifts for all nine sources detected with IRS. For three of them the redshifts were previously unknown. The median value of the redshift distribution is z~2.8 if we assume that the four IRS non-detections are at high redshift. The median for the IRS detections alone is z~2.7. Placing the IRS non-detections at similar redshift would require rest frame mid-IR obscuration larger than is seen in local ULIRGs. The rest frame mid-infrared spectra and mid- to far-infrared spectral energy distributions are consistent with those of local ultraluminous infrared galaxies, but scaled-up further in luminosity. The mid-infrared spectra support the scenario that submillimeter galaxies are sites of extreme star formation, rather than X-ray-obscured AGN, and represent a critical phase in the formation of massive galaxies., Comment: 31 pages, 4 figures. Accepted for publication in ApJ

We present a detailed analysis of the spatially-resolved properties of the lensed submillimeter galaxy SMMJ14011+0252 at z=2.56, combining deep near-infrared integral-field data obtained with SPIFFI on the VLT with other multi-wavelength data sets. The broad characteristics of SMMJ14011+0252 are in agreement with what is expected for the early evolution of local massive spheroidal galaxies. From continuum and line flux, velocity, and dispersion maps, we measure the kinematics, star-formation rates, gas densities, and extinction for individual subcomponents. The star formation intensity is similar to low-redshift ``maximal starbursts'', while the line fluxes and the dynamics of the emission line gas provide direct evidence for a starburst-driven wind with physical properties very similar to local superwinds. We also find circumstantial evidence for "self-regulated" star formation within J1. The relative velocity of the bluer companion J2 yields a dynamical mass estimate for J1 within about 20 kpc, M_dyn \sim 1\times 10^{11} M_sun. The relative metallicity of J2 is 0.4 dex lower than in J1n/s, suggesting different star formation histories. SED fitting of the continuum peak J1c confirms and substantiates previous suggestions that this component is a z=0.25 interloper. When removing J1c, the stellar continuum and H-alpha line emission appear well aligned spatially in two individual components J1n and J1s, and coincide with two kinematically distinct regions in the velocity map, which might well indicate a merging system. This highlights the close similarity between SMGs and ULIRGs, which are often merger-driven maximal starbursts, and suggests that the intrinsic mechanisms of star-formation and related feedback are similar to low-redshift strongly star-forming systems., Some of the figures changed from b/w to color

To explain the properties of the most massive low-redshift galaxies and the shape of their mass function, recent models of galaxy evolution include strong AGN feedback to complement starburst-driven feedback in massive galaxies. Using the near-infrared integral-field spectrograph SPIFFI on the VLT, we searched for direct evidence for such a feedback in the optical emission line gas around the z=2.16 powerful radio galaxy MRC1138-262, likely a massive galaxy in formation. The kpc-scale kinematics, with FWHMs and relative velocities L* galaxy within a few 10 to 100 Myrs, fast enough to preserve the observed [alpha/Fe] overabundance in massive galaxies at low redshift. Using simple arguments, it appears that feedback like that observed in MRC1138-262 may have sufficient energy to inhibit material from infalling into the dark matter halo and thus regulate galaxy growth as required in some recent models of hierarchical structure formation.

This paper addresses the origin of the far-infrared (FIR) continuum of QSOs, based on the Quasar and ULIRG Evolution Study (QUEST) of nearby QSOs and ULIRGs using observations with the Spitzer Space Telescope. For 27 Palomar-Green QSOs at z, Comment: 39 pages, 8 figures, accepted for publication in ApJ

We have performed a spectroscopic study of 7 Type 2 QSOs using the mid-infrared spectrometer IRS on board the Spitzer Space Telescope. These are (to our knowledge) the first mid-IR spectra of X-ray selected QSO2s taken. The objects have been selected according to their high intrinsic luminosities and column densities in the X-rays. Their spectra strongly differ from template spectra of Type 2 AGN at lower luminosities. They do not exhibit strong PAH dust emission features from circum-nuclear star forming regions, typical for lower luminosity Type 2 Seyfert galaxies or other previously used QSO2 templates, such as the (Ultra)luminous Infrared Galaxy ((U)LIRG) NGC 6240. They also do not show the ice and silicate absorption features of highly luminous but deeply embedded compact nuclei seen in some ULIRGs. Instead they reveal a relatively featureless, rising continuum similar to luminous Type 1 AGN. We also find evidence for a 10 micron silicate feature in emission. Models of dusty tori in the AGN unification scenario predict this only for Type 1 AGN. The ratio of the AGN continuum luminosity at 6 micron to the absorption corrected 2-10keV X-ray AGN luminosity is very similar to that found in Seyfert galaxies. X-ray selected QSO2s are thus characterized by powerful AGN in hosts with a luminosity due to star formation < 1e11 L_sol. The dominance of the AGN light in the mid-IR spectra of QSO2s together with their flatter spectral energy distributions (SEDs) places important constraints on models of the cosmic infrared background and of the star formation history of the universe.

International audience; We have observed a bright flare of Sgr A* in the near-infrared with the adaptive optics-assisted integral-field spectrometer SINFONI. Within the uncertainties, the observed spectrum is featureless and can be described by a power law. Our data suggest that the spectral index is correlated with the instantaneous flux and that both quantities experience significant changes within less than 1 hour. We argue that the near-infrared flares from Sgr A* are due to synchrotron emission of transiently heated electrons, the emission being affected by orbital dynamics and synchrotron cooling, both acting on timescales of ~20 minutes.

We present a deep 16.8 ks V-band image of the field of a candidate z=10 galaxy magnified by the foreground (z=0.25) cluster A1835. The image was obtained with FORS1 on VLT-Kueyen to test whether the V-band lies below the Lyman limit for this very high redshift candidate. A detection would unambiguously rule out that the source is at z=10. The 3sigma detection limit of the image in the area of the z=10 candidate is V_AB=28.0 mag in a 2 arcsec diameter aperture (about 3 times the seeing FWHM of 0.7 arcsec). No source at the position of the candidate galaxy is detected down to this limit. Formally, this is consistent with the V-band probing below the Lyman limit in the rest-frame of a z=10 source. However, given the recent non-detection of the object in a deep H-band exposure with NIRI on Gemini North down to H_AB=26.0 mag (3sigma in a 1.4 arcsec aperture) and concerns about the detection of the reported associated emission line, it may be possible that this source is spurious. We discuss several astrophysical possibilities to explain the puzzling nature of this source and find none of them compelling.

We discuss identifications for 18 sources from our MAMBO 1.2mm survey of the region surrounding the NTT Deep Field. We have obtained accurate positions from Very Large Array 1.4GHz interferometry and in a few cases IRAM mm interferometry, and have also made deep BVRIzJK imaging at ESO. We find thirteen 1.2mm sources associated with optical/near-infrared objects in the magnitude range K=19.0 to 22.5, while five are blank fields at K>22. The median redshift of the radio-identified mm sources is ~2.6 from the radio/mm estimator, and the median optical/near-infrared photometric redshifts for the objects with counterparts ~2.1. This suggests that those radio-identified mm sources without optical/near-infrared counterparts tend to lie at higher redshifts than those with optical/near-infrared counterparts. Compared to published identifications of objects from 850micron surveys of similar depth, the median K and I magnitudes of our counterparts are roughly two magnitudes fainter and the dispersion of I-K colors is less. Real differences in the median redshifts, residual mis-identifications with bright objects, cosmic variance, and small number statistics are likely to contribute to this significant difference, which also affects redshift measurement strategies. We discuss basic properties of the near-infrared/(sub)mm/radio spectral energy distributions of our galaxies and of interferometrically identified submm sources from the literature. From a comparison with submm objects with CO-confirmed spectroscopic redshifts we argue that roughly two thirds of the (sub)mm galaxies are at z>~2.5. This fraction is probably larger when including sources without radio counterparts. (abridged)

We have used the IRAM Plateau de Bure Interferometer to map CO(3-2) emission from the gravitationally lensed Lyman break galaxy MS1512-cB58. This is the first detection of a molecular emission line in any Lyman break system; its integrated intensity implies a total molecular gas mass of 6.6e9 Msun, while its width implies a dynamical mass of 1.0e10 csc^2i Msun (for a flat Lambda=0.7 cosmology). These estimates are in excellent concordance with nearly all parameters of the system measured at other wavelengths, and yield a consistent picture of past and future star formation with no obvious discrepancies requiring explanation by differential lensing. In particular, we find that the age and remaining lifetime of the current episode of star formation are likely to be similar; the surface densities of star formation and molecular gas mass are related by a Schmidt law; and the fraction of baryonic mass already converted into stars is sufficient to account for the observed enrichment of the interstellar medium to 0.4 Zsun. Barring substantial gas inflow or a major merger, the stars forming in the current episode will have mass and coevality at z=0 similar to those of a spiral bulge. Assuming cB58 is a typical Lyman break galaxy apart from its magnification, its global parameters suggest that the prescriptions for star formation used in some semi-analytic models of galaxy evolution require moderate revision, although the general prediction that gas mass fraction should increase with redshift is validated. [abridged], Comment: 41 pages, 6 figures, accepted by ApJ

We have discovered six galaxies with spectroscopically confirmed redshifts of 4.81.5 in the field, this photometric cut designed to select galaxies at z>4.8. The line fluxes range between 0.2 to 2.5 x 10^-17 ergs cm^-2 s^-1 indicating luminosities of around 10^42-43 ergs s^-1 for Ly-alpha and their high emission line equivalent widths suggest very young ages (~5.3 ionized. These galaxies are observed within several hundred Myr of the end of the epoch of reionization (z=6-7), with little time for the luminosity function to evolve. This, and the lack of detected quasars, imply that the bulk of the UV flux that reionized the universe came from faint galaxies with M_(1700 ang)>-21., Comment: 24 pages (latex, aas macros), 7 figures (12 postscript files), accepted for publication in the Astrophysical Main Journal

We study the heating mechanisms and Ly{\alpha} escape fractions of 35 Ly{\alpha} blobs (LABs) at z = 3.1 in the SSA22 field. Dust continuum sources have been identified in 11 of the 35 LABs, all with star formation rates (SFRs) above 100 Msun/yr. Likely radio counterparts are detected in 9 out of 29 investigated LABs. The detection of submm dust emission is more linked to the physical size of the Ly{\alpha} emission than to the Ly{\alpha} luminosities of the LABs. A radio excess in the submm/radio detected LABs is common, hinting at the presence of active galactic nuclei. Most radio sources without X-ray counterparts are located at the centers of the LABs. However, all X-ray counterparts avoid the central regions. This may be explained by absorption due to exceptionally large column densities along the line-of-sight or by LAB morphologies, which are highly orientation dependent. The median Ly{\alpha} escape fraction is about 3\% among the submm-detected LABs, which is lower than a lower limit of 11\% for the submm-undetected LABs. We suspect that the large difference is due to the high dust attenuation supported by the large SFRs, the dense large-scale environment as well as large uncertainties in the extinction corrections required to apply when interpreting optical data., Comment: 22 pages, 7 figures, ready for publication on ApJ

We have made a deep NICMOS imaging study of a sample of 5 z ~ 2 - 3 radio-quiet quasars with low absolute nuclear luminosities, and we have detected apparent host galaxies in all of these. Most of the hosts have luminosities approximately equal to present-day L*, with a range from 0.2 L* to about 4 L*. These host galaxies have magnitudes and sizes consistent with those of the Lyman break galaxies at similar redshifts and at similar rest wavelengths, but are about two magnitudes fainter than high-z powerful radio galaxies. The hosts of our high-z sample are comparable to or less luminous than the hosts of the low-z RQQs with similar nuclear absolute magnitudes. However, the high z galaxies are more compact than the hosts of the low z quasars, and probably have only 10 - 20% of the stellar mass of their low-z counterparts. Application of the M(bulge)/M(BH) relation found for present-day spheroids to the stellar masses implied for the high z host galaxies would indicate that they contain black holes with masses around 10^8 Msolar. Comparison to their nuclear magnitudes implies accretion rates that are near or at the Eddington limit. Although these high z hosts already contain supermassive black holes, the galaxies will need to grow significantly to evolve into present-day L* galaxies. These results are basically consistent with theoretical predictions for the hierarchical buildup of the galaxy host and its relation to the central supermassive black hole., 25 pages, 13 figures, accepted for publication in ApJ

Integral field spectroscopy in the K-band (1.9-2.4um) was performed on four IR-bright star clusters and the two nuclei in NGC 4038/4039 (``The Antennae''). Two of the clusters are located in the overlap region of the two galaxies, and together comprise ~25% of the total 15um and ~10% of the total 4.8 GHz emission from this pair of merging galaxies. The other two clusters, each of them spatially resolved into two components, are located in the northern galaxy, one in the western and one in the eastern loop of blue clusters. Comparing our analysis of Brgamma, CO band-heads, He I (2.058um), Halpha (from archival HST data), and V-K colors with stellar population synthesis models indicates that the clusters are extincted (A_V ~ 0.7 - 4.3 mags) and young, displaying a significant age spread (4-13 Myrs). The starbursts in the nuclei are much older (65 Myrs), with the nucleus of NGC 4038 displaying a region of recent star formation northward of its K-band peak. Using our derived age estimates and assuming the parameters of the IMF (Salpeter slope, upper mass cut-off of 100 M_sun, Miller-Scalo between 1 M_sun and 0.1 M_sun), we find that the clusters have masses between 0.5 and 5 * 10^6M_sun.

We report the detection of the Diffuse Interstellar Bands (``$DIBs$'') in the optical spectra of seven far-infrared-selected starburst galaxies. The $\lambda$6283.9 \AA and $\lambda$ 5780.5 \AA features are detected with equivalent widths of $\sim$ 0.4 to 1 \AA and 0.1 to 0.6 \AA respectively. In the two starbursts with the highest quality spectra (M82 and NGC2146), four other weaker $DIBs$ at $\lambda$ 5797.0 \AA, 6010.1 \AA, 6203.1 \AA, and 6613.6 \AA are detected with equivalent widths of $\sim$ 0.1 \AA. The region over which the $DIBs$ can be detected ranges from $\sim$ 1 kpc in the less powerful starbursts, to several kpc in the more powerful ones. The gas producing the $DIBs$ is more kinematically quiescent on-average than the gas producing the strongly-blueshifted $NaI\lambda\lambda$5890,5896 absorption in the same starbursts. We show that the $DIBs$ in these intense starbursts are remarkably similar to those in our Galaxy: the relative strengths of the features detected are similar, and the equivalent widths follow the same dependence as Galactic $DIBs$ on $E(B-V)$ and $NaI$ column density. While the ISM in starbursts is heated by a photon and cosmic ray bath that is $\sim$ 10$^3$ times more intense than in the diffuse ISM of the Milky Way, the gas densities and pressures are also correspondingly larger in starbursts. This ``homology'' may help explain the strikingly similar $DIB$ properties., Comment: 18 pages, 5 figures, in press in ApJ

By means of KPNO 2.1 m J-, H-, and K-band data on samples of gigahertz peaked spectrum (GPS, radio size 20 kpc), we have constructed mean broadband spectral energy distributions (SEDs) for each of the object classes and compared these with synthetic SEDs. The host galaxies of the compact and extended radio sources appear to have similar near-IR broadband SEDs, implying similar stellar populations and active galactic nuclei (AGNs). These SEDs are best fitted with a mean metallicity ~solar and fairly old stellar populations implying formation redshifts of 5. Redshift evolution of the sources is consistent with a passive evolutionary model. In addition, an extra IR component is needed that can be modeled by emission from dust at a temperature of 900-1300 K, perhaps produced by the putative obscuring torus. All of our results are consistent with the hypothesis that the GPS and CSS are the progenitors of the large-scale 3CR radio galaxies.

In this paper we present the data from our J-, H-, and K-band imaging program of powerful radio sources. The sample consists of radio galaxies classified as either compact ( 20 kpc) radio sources. Source images, photometry results, radial profiles, and profile fits are presented. Source decomposition of total flux into point-spread function (PSF) and extended contributions indicate all sources are nucleated to roughly the same level at J, H, and K bands, arguing against the notion that the suspected higher nuclear absorption in the most compact class would render those sources relatively more nucleated at the longer wavelengths. No significant differences in near-IR host galaxy properties have been found. In a companion paper we discuss color-color properties, comparisons with predictions made by a stellar synthesis code, and near-IR broadband spectral properties.

The broadband X-ray spectrum of the prototypical starburst galaxy M82 is very complex. At least three spectral components are required to fit the combined ROSAT and ASCA spectrum in the 0.1--10 keV range. The observed X-ray flux in this band is dominated by a hard Gamma = 1.7, heavily absorbed power law component which originates in the nucleus and near-nuclear disk of the galaxy. Among the candidates for the origin of this hard X-ray emission, the most plausible appears to be inverse-Compton scattered emission from the interaction of M82's copious infrared photon flux with supernova-generated relativistic electrons. The measured intrinsic luminosity of the power law component agrees closely with calculations of the expected inverse-Compton luminosity. Moreover, the radio and X-ray emission in the nucleus of M82 have the same spectral slope, which should be the case if both types of emission are non- thermal and are associated with a common population of electrons. The other two spectral components, thermal plasmas with characteristic temperatures kT = 0.6 and 0.3 keV, are associated with the star formation and starburst-driven wind in M82. The warmer thermal component is heavily absorbed as well and must also originate in the central region of the galaxy. The softer thermal component, however, is not absorbed, and is likely to represent the X-ray emission which extends along M82's minor axis. The amount of absorption required in the three-component model suggests that the intrinsic luminosity of M82 in the 0.1--10 keV band is about four times greater than its observed luminosity of 4 x 10(40) ergs/s., Comment: 23 pages, TeX, seven postscript figures, three tables Accepted for the March 20, 1997 issue of ApJ

We report the first detection of ~10 and ~18 micron silicate dust emissions in a low-luminosity active galactic nucleus (AGN), obtained in Spitzer-IRS 7-37 micron spectroscopy of the Type 1 LINER galaxy NGC3998. Silicate emissions in AGN have only recently been detected in several quasars. Our detection counters suggestions that silicate emissions are present only in the most luminous AGN. The silicate features may be signatures of a dusty ``obscuring torus'' viewed face-on as postulated for Type 1 AGN. However, the apparently cool (~200 K) dust is inconsistent with theoretical expectations of much hotter torus walls. Furthermore, not all Type 1 objects are silicate emission sources. Alternatively, the silicate emission may originate in dust not directly associated with a torus. We find that the long-wavelength (>20 micron) tail of the emission in NGC3998 is significantly weaker than in the sample of bright QSOs recently presented by Hao et al. The 10 micron profile in our NGC3998 spectrum is inconsistent with ``standard'' silicate ISM dust. This may indicate differences in the dust composition, grain size distribution, or degree of crystallization. The differences between NGC3998, QSOs, and Galactic templates suggest that there are significant environmental variations., Accepted by ApJL, 10 pages, 2 eps figures

Utilizing a large sample of infrared-selected starburst galaxies having optical images and long-slit spectra, we explore the interrelationships between the properties of starbursts and relate these properties to those of the "host" galaxy. We find that the half-light radius of the Hα-emitting region (re, Hα) enters into several correlations that suggest it is physically related to the actual starburst radius. Most suggestively, the effective IR surface brightness (LIR/πr2e, Hα) correlates strongly with the far-IR color temperature. This can be reproduced roughly with an idealized model of a surrounding dust screen whose far-IR emissivity is determined by the local energy density of UV starburst light. Typical values for re, Hα are a few hundred pc to a few kpc (with the Ha emission being significantly more compact than the red starlight). This confirms the "circumnuclear" scales of typical starbursts. We show also that starbursts seem to obey a limiting IR surface brightness of about 1011 L☉ kpc2, corresponding to a maximum star formation rate of about 20 M☉ yr–1 kpc2 for a normal initial mass function. We argue that this upper limit suggests that starbursts are self-regulating in some way.We show that most of these galaxies have relatively normal, symmetric rotation curves. This implies that the galactic disk need not suffer severe dynamical damage in order to "fuel" a typical starburst. We show also that the starbursts occur preferentially in the inner region of solid-body rotation. This may reflect both bar-driven inflow of gas to the region between the inner Lindblad resonances and the dominance of gravitational instability over tidal shear in this region. Most of the starbursts reside in galaxies with rotation speeds of 120--200 km s–1 (compared to 220 km s–1 for a fiducial L* galaxy like the Milky Way). The lack of a correlation between galaxy rotation speed and starburst luminosity means that even relatively modest galaxies (masses ~10% of the Milky Way) can host powerful starbursts. We argue on the basis of causality that the internal velocity dispersion in a starburst sets an upper limit to the star formation rate. The most extreme starbursts approach this limit, but most are well below. Finally, we show that the relative narrowness of the nuclear emission lines in starbursts (relative to the galaxy rotation speed) arises because the gas in the nuclear "bin" usually does not sample fully the solid-body part of the rotation curve. The narrow lines do not necessarily imply that the starburst is not in dynamical equilibrium.