Mode Of Star Formation Research Articles

The stellar initial mass function in red-sequence galaxies: 1-μm spectroscopy of Coma cluster galaxies with Subaru/FMOS

To investigate possible variations in the stellar initial mass function (IMF) in red-sequence galaxies, we have obtained infrared spectroscopy with Subaru/FMOS for a sample of 92 red-sequence galaxies in the Coma cluster. Velocity dispersions, ages and element abundances for these galaxies have been previously determined from optical data. By stacking the FMOS spectra in the rest frame, removing sky-subtraction residuals and other artefacts fixed in the observed frame, we derive composite spectra in the 9600-10500 Angstrom range for galaxies grouped according to their velocity dispersion or Mg/Fe ratio. We measure the Wing-Ford band of FeH and a new index centred on a CaI line at 10345 Angstrom; these features are strong in cool dwarf stars, and hence reflect the form of the IMF at low mass (<0.5M_sun) The CaI line, unlike the Wing-Ford band and other `classical' IMF indicators (NaI doublet, CaII triplet), is unaffected by the abundance of sodium. We compare the measured indices against predictions from spectral synthesis models matched to the element abundances estimated from the optical data. Binning galaxies by velocity dispersion, we find that both the Wing-Ford and CaI index measurements are best reproduced by models with the Salpeter IMF. There is no clear evidence for an increase in dwarf-star content with velocity dispersion. Binning the observed galaxies instead by Mg/Fe ratio, the behaviour of both indices implies a trend of IMF from Chabrier-like, at abundance ratios close to solar, to Salpeter or heavier for highly alpha-enhanced populations. At face value, this suggests that the IMF depends on the mode of star formation, with intense rapid star-bursts generating a larger population of low-mass stars.

A UNIFIED EMPIRICAL MODEL FOR INFRARED GALAXY COUNTS BASED ON THE OBSERVED PHYSICAL EVOLUTION OF DISTANT GALAXIES

We reproduce the mid-infrared to radio galaxy counts with a new empirical model based on our current understanding of the evolution of main-sequence (MS) and starburst (SB) galaxies. We rely on a simple Spectral Energy Distribution (SED) library based on Herschel observations: a single SED for the MS and another one for SB, getting warmer with redshift. Our model is able to reproduce recent measurements of galaxy counts performed with Herschel, including counts per redshift slice. This agreement demonstrates the power of our 2 Star-Formation Modes (2SFM) decomposition for describing the statistical properties of infrared sources and their evolution with cosmic time. We discuss the relative contribution of MS and SB galaxies to the number counts at various wavelengths and flux densities. We also show that MS galaxies are responsible for a bump in the 1.4 GHz radio counts around 50 {\mu}Jy. Material of the model (predictions, SED library, mock catalogs...) is available online at http://irfu.cea.fr/Sap/Phocea/Page/index.php?id=537.

Modes of clustered star formation

The realization that most stars form in clusters, raises the question of whether star/planet formation are influenced by the cluster environment. The stellar density in the most prevalent clusters is the key factor here. Whether dominant modes of clustered star formation exist is a fundamental question. Using near-neighbour searches in young clusters Bressert et al. (2010) claim this not to be the case and conclude that star formation is continuous from isolated to densely clustered. We investigate under which conditions near-neighbour searches can distinguish between different modes of clustered star formation. Near-neighbour searches are performed for model star clusters investigating the influence of the combination of different cluster modes, observational biases, and types of diagnostic and find that the cluster density profile, the relative sample sizes, limitations in observations and the choice of diagnostic method decides whether modelled modes of clustered star formation are detected. For centrally concentrated density distributions spanning a wide density range (King profiles) separate cluster modes are only detectable if the mean density of the individual clusters differs by at least a factor of ~65. Introducing a central cut-off can lead to underestimating the mean density by more than a factor of ten. The environmental effect on star and planet formation is underestimated for half of the population in dense systems. A analysis of a sample of cluster environments involves effects of superposition that suppress characteristic features and promotes erroneous conclusions. While multiple peaks in the distribution of the local surface density imply the existence of different modes, the reverse conclusion is not possible. Equally, a smooth distribution is not a proof of continuous star formation, because such a shape can easily hide modes of clustered star formation (abridged)

Mahalo-Subaru: Mapping Star Formation at the Peak Epoch of Massive Galaxy Formation

AbstractMAHALO-Subaru (MApping HAlpha and Lines of Oxygen with Subaru) is our on-going large programme which aims to investigate how the star forming activities in galaxies are propagated as a function of time, mass, and environment. We are targeting 10 clusters and proto-clusters at 0.4&lt;z&lt;2.6, and two general fields (GOODS-N and SXDF-CANDELS) with Suprime-Cam and MOIRCS by utilizing our unique sets of narrow-band filters. The narrow-band imaging can map out star forming galaxies with the redshifted Halpha and/or [OII] emission lines from our targets, and thus providing relatively unbiased views of star forming activities across time and environment. We have almost completed narrow-band imaging of our targets, and found that star forming activity is very high even in the proto-cluster cores (z≳1.5), and that the peak of star formation is shifted outwards with time, indicating the inside-out formation of clusters. Moreover, we have identified many “red” emitters especially in high density regions at z&gt;2, which suggests that the mode of star formation and/or the activation of AGN are dependent on environment, and thus holding the key to the environmental effects at the early stage of cluster galaxies formation and evolution.

Probing the star–formation modes in merging galaxies

AbstractMerging systems at low redshift provide the unique opportunity to study the processes related to star formation in a variety of environments that presumably resemble those seen at higher redshifts. Previous studies of distant starbursting galaxies suggest that stars are born in turbulent gas, with a higher efficiency than in MW-like spirals. We have investigated in detail the turbulent-driven regime of star-formation in nearby colliding galaxies combining high resolution VLA B array Hi maps and UV GALEX observations. With these data, we could check predictions of our state-of-the-art simulations of mergers, such as the global sharp increase of the fraction of dense gas, as traced by the SFR, with respect to the diffuse gas traced by Hi during the merging stage, following the increased velocity dispersion of the gas. We present here initial results obtained studying the SFR-Hi relation at 4.5 kpc resolution. We determined SFR/Hi mass ratios that are higher in the external regions of mergers than in the outskirts of isolated spirals, though both environments are Hi dominated. SFR/Hi increases towards the central regions following the decrease of the atomic gas fraction and possibly the increased star–formation efficiency. These results need to be checked with a larger sample of systems and on smaller spatial scales. This is the goal of the on-going Chaotic THINGS project that ultimately will allow us to determine why starbursting galaxies deviate from the Kennicutt-Schmidt relation between SFR density and gas surface density.

Star Formation &amp; Molecular Gas over Cosmic Time

AbstractRecent observations have revealed the existence of a ‘main sequence’ of star-forming galaxies out to high redshift. While the majority of star-forming galaxies are observed to be close to this relation between star formation rate (SFR) and stellar mass, a smaller subset of the population – so-called ‘starbursts’ – displays specific star-formation rates and star-formation efficiencies that exceed those of normal (main-sequence) galaxies by up to an order of magnitude. A large degree of homogeneity and similarity has been observed for the properties of the population of normal galaxies across a broad redshift range, including a narrow correlation between their CO luminosity (hence gas content) and IR luminosity and an almost invariable IR SED getting warmer with redshift, while starburst galaxies display systematically different properties. This can be used to devise a simple description of the evolution of the star-forming galaxy population since z ~ 2 and, with a higher degree of uncertainty, even further back in time, in a scheme that we dub two star formation mode framework (2-SFM). We show how this can successfully reproduce the shape of the IR luminosity function of galaxies as a function of redshifts, and the IR number counts. Furthermore, we can link the cosmic evolution of the sSFR of main-sequence galaxies to the evolution of the molecular fuel reservoir and to derive estimates of the molecular gas mass functions of star-forming galaxies that are based on their empirically measured gas properties rather than simulations or semi-analytical modelling. We also infer the evolution of the cosmic abundance of molecular gas and briefly discuss its expected observational signature by molecular line emission, the CO luminosity function.

THE STAR FORMATION HISTORY AND CHEMICAL EVOLUTION OF STAR-FORMING GALAXIES IN THE NEARBY UNIVERSE

We have determined the O/H and N/O of a sample of 122751 SFGs from the DR7 of the SDSS. For all these galaxies we have also determined their morphology and their SFH using the code STARLIGHT. The comparison of the chemical abundance with the SFH allows us to describe the chemical evolution in the nearby universe (z < 0.25) in a manner which is consistent with the formation of their stellar populations and morphologies. A 45% of the SFGs in our sample show an excess of abundance in nitrogen relative to their metallicity. We also find this excess to be accompanied by a deficiency of oxygen, which suggests that this could be the result of effective starburst winds. However, we find no difference in the mode of star formation of the nitrogen rich and nitrogen poor SFGs. Our analysis suggests they all form their stars through a succession of bursts of star formation extended over a few Gyr period. What produces the chemical differences between these galaxies seems therefore to be the intensity of the bursts: the galaxies with an excess of nitrogen are those that are presently experiencing more intense bursts, or have experienced more intense bursts in their past. We also find evidence relating the chemical evolution process to the formation of the galaxies: the galaxies with an excess of nitrogen are more massive, have more massive bulges and earlier morphologies than those showing no excess. As a possible explanation we propose that the lost of metals consistent with starburst winds took place during the formation of the galaxies, when their potential wells were still building up, and consequently were weaker than today, making starburst winds more efficient and independent of the final mass of the galaxies. In good agreement with this interpretation, we also find evidence consistent with downsizing, according to which the more massive SFGs formed before the less massive ones.

THE FIRST GALAXIES: ASSEMBLY WITH BLACK HOLE FEEDBACK

We study how the first galaxies were assembled under feedback from the accretion onto a central black hole (BH) that is left behind by the first generation of metal-free stars through self-consistent, cosmological simulations. X-ray radiation from the accretion of gas onto BH remnants of Population III (Pop III) stars, or from high-mass X-ray binaries (HMXBs), again involving Pop III stars, influences the mode of second generation star formation. We track the evolution of the black hole accretion rate and the associated X-ray feedback starting with the death of the Pop III progenitor star inside a minihalo and following the subsequent evolution of the black hole as the minihalo grows to become an atomically cooling galaxy. We find that X-ray photoionization heating from a stellar-mass BH is able to quench further star formation in the host halo at all times before the halo enters the atomic cooling phase. X-ray radiation from a HMXB, assuming a luminosity close to the Eddington value, exerts an even stronger, and more diverse, feedback on star formation. It photoheats the gas inside the host halo, but also promotes the formation of molecular hydrogen and cooling of gas in the intergalactic medium and in nearby minihalos, leading to a net increase in the number of stars formed at early times. Our simulations further show that the radiative feedback from the first BHs may strongly suppress early BH growth, thus constraining models for the formation of supermassive BHs.

How to distinguish starbursts and quiescently star-forming galaxies: the ‘bimodal’ submillimetre galaxy population as a case study

In recent work (arXiv:1101.0002) we have suggested that the high-redshift (z ~ 2-4) bright submillimetre galaxy (SMG) population is heterogeneous, with major mergers contributing both at early stages, where quiescently star-forming discs are blended into one submm source (`galaxy-pair SMGs'), and late stages, where mutual tidal torques drive gas inflows and cause strong starbursts. Here we combine hydrodynamic simulations of major mergers with 3-D dust radiative transfer calculations to determine observational diagnostics that can distinguish between quiescently star-forming SMGs and starburst SMGs via integrated data alone. We fit the far-IR SEDs of the simulated galaxies with the optically thin single-temperature modified blackbody, the full form of the single-temperature modified blackbody, and a power-law temperature-distribution model. The effective dust temperature, T_dust, and power-law index of the dust emissivity in the far-IR, \beta, derived can significantly depend on the fitting form used, and the intrinsic \beta\ of the dust is not recovered. However, for all forms used here, there is a T_dust above which almost all simulated galaxies are starbursts, so a T_dust cut is very effective at selecting starbursts. Simulated merger-induced starbursts also have higher L_IR/M_gas and L_IR/L_FUV than quiescently star-forming galaxies and lie above the star formation rate-stellar mass relation. These diagnostics can be used to test our claim that the SMG population is heterogeneous and to observationally determine what star formation mode dominates a given galaxy population. We comment on applicability of these diagnostics to ULIRGs that would not be selected as SMGs. These `hot-dust ULIRGs' are typically starburst galaxies lower in mass than SMGs, but they can also simply be SMGs observed from a different viewing angle.

CONSTRAINTS ON FEEDBACK PROCESSES DURING THE FORMATION OF EARLY-TYPE GALAXIES

Galaxies are found to obey scaling relations between a number of observables. These relations follow different trends at the low- and the high-mass ends. The processes driving the curvature of scaling relations remain uncertain. In this letter, we focus on the specific family of early-type galaxies, deriving the star formation histories of a complete sample of visually classified galaxies from SDSS-DR7 over the redshift range 0.01<z<0.025, covering a stellar mass interval from 10^9 to 3 x 10^11 Msun. Our sample features the characteristic "knee" in the surface brightness vs. mass distribution at Mstar~3 x 10^10 Msun. We find a clear difference between the age and metallicity distributions of the stellar populations above and beyond this knee, which suggests a sudden transition from a constant, highly efficient mode of star formation in high-mass galaxies, gradually decreasing towards the low-mass end of the sample. At fixed mass, our early-type sample is more efficient in building up the stellar content at early times in comparison to the general population of galaxies, with half of the stars already in place by redshift z~2 for all masses. The metallicity-age trend in low-mass galaxies is not compatible with infall of metal-poor gas, suggesting instead an outflow-driven relation.

On turbulent fragmentation and the origin of the stellar IMF

Two varieties of the universal stellar initial mass function (IMF) viz., the Kroupa and the Chabrier IMF, have emerged over the last decade to explain the observed distribution of stellar masses. The possibility of the universal nature of the stellar IMF leads us to the interesting prospect of a universal mode of star-formation. It is well-known that turbulent fragmentation of gas in the interstellar medium produces a lognormal distribution of density which is further reflected by the mass-function for clumps at low and intermediate masses. Stars condense out of unstable clumps through a complex interplay between a number of dynamic processes which must be accounted for when tracing the origin of the stellar IMF. In the present work, applying the theory of gravitational fragmentation we first derive the mass function (MF) for clumps. Then a core mass function (CMF) is derived by allowing the clumps to fragment, having subjected each one to a random choice of gas temperature. Finally, the stellar IMF is derived by applying a random core-to-star conversion efficiency, ϵ, in the range of 5–15% to each CMF. We obtain a power-law IMF that has exponents within the error-bars on the Kropua IMF. This derived IMF is preceded by a similar core mass function which suggests, gravoturbulent fragmentation plays a key role in assembling necessary conditions that relate the two mass-functions. In this sense the star-formation process, at least at low redshifts where gas cooling is efficient, is likely to be universal. We argue that the observed knee in the CMF and the stellar IMF may alternatively be interpreted in terms of the characteristic temperature at which gas in potential star-forming clouds is likely to be found. Our results also show that turbulence in star-forming clouds is probably driven on large spatial scales with a power-spectrum steeper than Kolmogorov-type.

Contribution from star-forming galaxies to the cosmic gamma-ray background radiation

The origin of extragalactic gamma-ray background (EGRB) has been discussed for a long time and various sources have been discussed as possible contributors to EGRB, such as active galactic nuclei, star-forming galaxies, galaxy clusters, structure formation, and dark matter annihilation. We present a new theoretical calculation of the contribution to the EGRB from star-forming galaxies based on a state-of-the-art model of hierarchical galaxy formation that is in quantitative agreement with a variety of observations of local galaxies as well as high redshifts. Gamma-ray luminosity and spectrum of galaxies are related to star formation rate, gas mass, and star formation mode (quiescent or starburst) of model galaxies using latest observed data of nearby galaxies. Our standard model predicts ∼ 14% contribution from star-forming galaxies to the total EGRB flux recently reported by the Fermi Gamma-Ray Space Telescope. The predicted number of nearby galaxies detectable by Fermi is consistent with the observation. The sum of the known contributions from AGNs and star-forming galaxies can explain a large part of EGRB, with a remarkable agreement between the predicted model spectrum and observation.

THE STAR FORMATION HISTORY OF LEO T FROMHUBBLE SPACE TELESCOPEIMAGING

We present the star formation history (SFH) of the faintest known star-forming galaxy, Leo T, based on imaging taken with the Hubble Space Telescope (HST) Wide Field Planetary Camera 2 (WFPC2). The HST/WFPC2 color-magnitude diagram (CMD) of Leo T is exquisitely deep, extending ~ 2 magnitudes below the oldest main sequence turnoff, permitting excellent constraints on star formation at all ages. We use a maximum likelihood CMD fitting technique to measure the SFH of Leo T assuming three different sets of stellar evolution models: Padova (solar-scaled metallicity) and BaSTI (both solar-scaled and alpha-enhanced metallicities). The resulting SFHs are remarkably consistent at all ages, indicating that our derived SFH is robust to the choice of stellar evolution model. From the lifetime SFH of Leo T, we find that 50% of the total stellar mass formed prior to z ~ 1 (7.6 Gyr ago). Subsequent to this epoch, the SFH of Leo T is roughly constant until the most recent ~ 25 Myr, where the SFH shows an abrupt drop. This decrease could be due to a cessation of star formation or stellar initial mass function sampling effects, but we are unable to distinguish between the two scenarios. Overall, our measured SFH is consistent with previously derived SFHs of Leo T. However, the HST-based solution provides improved age resolution and reduced uncertainties at all epochs. The SFH, baryonic gas fraction, and location of Leo T are unlike any of the other recently discovered faint dwarf galaxies in the Local Group, and instead bear strong resemblance to gas-rich dwarf galaxies (irregular or transition), suggesting that gas-rich dwarf galaxies may share common modes of star formation over a large range of stellar mass (~ 10^5-10^9 Msun).

AHerschelview of the far-infrared properties of submillimetre galaxies

We study a sample of 61 submillimetre galaxies (SMGs) selected from ground-based surveys, with known spectroscopic redshifts and observed with Herschel as part of the PACS Evolutionary Probe (PEP) and the Herschel Multi-tiered Extragalactic Survey (HerMES) key programmes. We use the broad far-infrared wavelength coverage (100-600um) provided by the combination of PACS and SPIRE observations. Using a power-law temperature distribution model to derive infrared luminosities and dust temperatures, we measure a dust emissivity spectral index for SMGs of beta=2.0+/-0.2. Our results unveil the diversity of the SMG population. Some SMGs exhibit extreme infrared luminosities of ~10^13 Lsun and relatively warm dust components, while others are fainter (~10^12 Lsun) and are biased towards cold dust temperatures. The extreme infrared luminosities of some SMGs (LIR>10^12.7 Lsun, 26/61 systems) imply SFRs of >500Msun yr^-1. Such high SFRs are difficult to reconcile with a secular mode of star formation, and may instead correspond to a merger-driven stage in the evolution of these galaxies. Another observational argument in favour of this scenario is the presence of dust temperatures warmer than that of SMGs of lower luminosities (~40K as opposed to ~25K), consistent with observations of local ULIRGs triggered by major mergers and with results from hydrodynamic simulations of major mergers combined with radiative transfer calculations. Luminous SMGs are also offset from normal star-forming galaxies in the stellar mass-SFR plane, suggesting that they are undergoing starburst events with short duty cycles, compatible with the major merger scenario. On the other hand, a significant fraction of the low infrared luminosity SMGs have cold dust temperatures, are located close to the main sequence of star formation, and thus might be evolving through a secular mode of star formation. [abridged]

GALAXY STRUCTURE AND MODE OF STAR FORMATION IN THE SFR-MASS PLANE FROMz∼ 2.5 TOz∼ 0.1

We analyze the dependence of galaxy structure (size and Sersic index) and mode of star formation (\Sigma_SFR and SFR_IR/SFR_UV) on the position of galaxies in the SFR versus Mass diagram. Our sample comprises roughly 640000 galaxies at z~0.1, 130000 galaxies at z~1, and 36000 galaxies at z~2. Structural measurements for all but the z~0.1 galaxies were based on HST imaging, and SFRs are derived using a Herschel-calibrated ladder of SFR indicators. We find that a correlation between the structure and stellar population of galaxies (i.e., a 'Hubble sequence') is already in place since at least z~2.5. At all epochs, typical star-forming galaxies on the main sequence are well approximated by exponential disks, while the profiles of quiescent galaxies are better described by de Vaucouleurs profiles. In the upper envelope of the main sequence, the relation between the SFR and Sersic index reverses, suggesting a rapid build-up of the central mass concentration in these starbursting outliers. We observe quiescent, moderately and highly star-forming systems to co-exist over an order of magnitude or more in stellar mass. At each mass and redshift, galaxies on the main sequence have the largest size. The rate of size growth correlates with specific SFR, and so does \Sigma_SFR at each redshift. A simple model using an empirically determined SF law and metallicity scaling, in combination with an assumed geometry for dust and stars is able to relate the observed \Sigma_SFR and SFR_IR/SFR_UV, provided a more patchy dust geometry is assumed for high-redshift galaxies.

THE STAR FORMATION HISTORY OF THE MILKY WAY'S NUCLEAR STAR CLUSTER

We present spatially resolved imaging and integral field spectroscopy data for 450 cool giant stars within 1\,pc from Sgr\,A*. We use the prominent CO bandheads to derive effective temperatures of individual giants. Additionally we present the deepest spectroscopic observation of the Galactic Center so far, probing the number of B9/A0 main sequence stars ($2.2-2.8\,M_\odot$) in two deep fields. From spectro-photometry we construct a Hertzsprung-Russell diagram of the red giant population and fit the observed diagram with model populations to derive the star formation history of the nuclear cluster. We find that (1) the average nuclear star-formation rate dropped from an initial maximum $\sim10$\,Gyrs ago to a deep minimum 1-2\,Gyrs ago and increased again during the last few hundred Myrs, and (2) that roughly 80% of the stellar mass formed more than 5\,Gyrs ago; (3) mass estimates within $\rm R\sim1\,pc$ from Sgr\,A* favor a dominant star formation mode with a 'normal' Chabrier/Kroupa initial mass function for the majority of the past star formation in the Galactic Center. The bulk stellar mass seems to have formed under conditions significantly different from the young stellar disks, perhaps because at the time of the formation of the nuclear cluster the massive black hole and its sphere of influence was much smaller than today.

STAR CLUSTER COMPLEXES AND THE HOST GALAXY IN THREE H II GALAXIES: Mrk 36, UM 408, AND UM 461

We present a stellar population study of three HII galaxies (Mrk 36, UM 408, and UM 461) based on the analysis of new ground-based high resolution near-infrared J, H and Kp broad-band and Br narrow-band images obtained with Gemini/NIRI. We identify and determine relative ages and masses of the elementary star clusters and/or star cluster complexes of the starburst regions in each of these galaxies by comparing the colors with evolutionary synthesis models that include the contribution of stellar continuum, nebular continuum and emission lines. We found that the current star cluster formation efficiency in our sample of low luminosity HII galaxies is ~10%. Therefore, most of the recent star formation is not in massive clusters. Our findings seem to indicate that the star formation mode in our sample of galaxies is clumpy, and that these complexes are formed by a few massive star clusters with masses > 10^4 Mo. The age distribution of these star cluster complexes shows that the current burst started recently and likely simultaneously over short time scales in their host galaxies, triggered by some internal mechanism. Finally, the fraction of the total cluster mass with respect to the low surface brightness (or host galaxy) mass, considering our complete range in ages, is less than 1%.

GOODS- HERSCHEL : GAS-TO-DUST MASS RATIOS AND CO-TO-H 2 CONVERSION FACTORS IN NORMAL AND STARBURSTING GALAXIES AT HIGH- z

We explore the gas-to-dust mass ratio (G/D) and the CO luminosity-to-Mgas conversion factor (a_co) of two well studied galaxies in the GOODS-N field, that are expected to have different star forming modes, the starburst GN20 at z=4.05 and the normal star-forming galaxy BzK-21000 at z=1.52. Detailed sampling is available for their Rayleigh-Jeans emission via ground based mm interferometry (1.1-6.6mm) along with Herschel, PACS and SPIRE data that probe the peak of their infrared emission. Using the physically motivated Draine & Li (2007) models, as well as a modified black body function, we measure the dust mass (Md) of the sources and find 2.0^{+0.7}_{-0.6} x 10^{9} Msun for GN20 and 8.6^{+0.6}_{-0.9} x 10^{8} Msun for BzK-21000. The addition of mm data reduces the uncertainties of the derived Md by a factor of ~2, allowing the use of the local G\D vs metallicity relation to place constraints on the a_co values of the two sources. For GN20 we derive a conversion factor of a_co < 1.0 Msun pc^{-2}(K km s^{-1})^{-1}, consistent with that of local ULIRGs, while for BzK-21000 we find a considerably higher value, a_co ~4.0 Msun pc^{-2}(K km s^{-1})^{-1}, in agreement with an independent kinematic derivation reported previously. The implied star formation efficiency is ~25 Lsun/Msun for BzK-21000, a factor of ~5-10 lower than that of GN20. The findings for these two sources support the existence of different disk-like and starburst star-formation modes in distant galaxies, although a larger sample is required to draw statistically robust results

The WiggleZ Dark Energy Survey: high-resolution kinematics of luminous star-forming galaxies

We report evidence of ordered orbital motion in luminous star-forming galaxies at z~1.3. We present integral field spectroscopy (IFS) observations, performed with the OH Suppressing InfraRed Imaging Spectrograph (OSIRIS) system, assisted by laser guide star adaptive optics on the Keck telescope, of 13 star-forming galaxies selected from the WiggleZ Dark Energy Survey. Selected via ultraviolet and [OII] emission, the large volume of the WiggleZ survey allows the selection of sources which have comparable intrinsic luminosity and stellar mass to IFS samples at z>2. Multiple 1-2 kpc size sub-components of emission, or 'clumps', are detected within the Halpha spatial emission which extends over 6-10 kpc in 4 galaxies, resolved compact emission (r<3 kpc) is detected in 5 galaxies, and extended regions of Halpha emission are observed in the remaining 4 galaxies. We discuss these data in the context of different snapshots in a merger sequence and/or the evolutionary stages of coalescence of star-forming regions in an unstable disk. We find evidence of ordered orbital motion in galaxies as expected from disk models and the highest values of velocity dispersion (\sigma>100 km/s) in the most compact sources. This unique data set reveals that the most luminous star-forming galaxies at z>1 are gaseous unstable disks indicating that a different mode of star formation could be feeding gas to galaxies at z>1, and lending support to theories of cold dense gas flows from the intergalactic medium.

THE LESSER ROLE OF STARBURSTS IN STAR FORMATION AT z = 2

Two main modes of star formation are know to control the growth of galaxies: a relatively steady one in disk-like galaxies, defining a tight star formation rate (SFR)-stellar mass sequence, and a starburst mode in outliers to such a sequence which is generally interpreted as driven by merging. Such starburst galaxies are rare but have much higher SFRs, and it is of interest to establish the relative importance of these two modes. PACS/Herschel observations over the whole COSMOS and GOODS-South fields, in conjunction with previous optical/near-IR data, have allowed us to accurately quantify for the first time the relative contribution of the two modes to the global SFR density in the redshift interval 1.5<z<2.5, i.e., at the cosmic peak of the star formation activity. The logarithmic distributions of galaxy SFRs at fixed stellar mass are well described by Gaussians, with starburst galaxies representing only a relatively minor deviation that becomes apparent for SFRs more than 4 times higher than on the main sequence. Such starburst galaxies represent only 2% of mass-selected star forming galaxies and account for only 10% of the cosmic SFR density at z~2. Only when limited to SFR>1000M(sun)/yr, off-sequence sources significantly contribute to the SFR density (46+/-20%). We conclude that merger-driven starbursts play a relatively minor role for the formation of stars in galaxies, whereas they may represent a critical phase towards the quenching of star formation and morphological transformation in galaxies.