Starburst galaxies – An observer’s view

Starbursts are the most efficient producers of metals in the Universe at low redshifts. They produce enough energy to driveoutflows of material from their disks.This makes them important objects to study in order to understand the chemical evolution not only of the interstellar medium (ISM) in the starburst galaxies themselves, but also of the intergalactic medium (IGM) in their vicinity. However, several key quantities of starbursts that are neededas input to models of their ISM are still ill-constrained. Some of these critical parameters are e.g. the metalabundances of hot ionized gas, the ionization state ofwarm ionized gas, the amount of energy deposited intothe ambient by a starburst, the efficiency of itsconversion into mechanical energy and thus the totalkinetic energy of the star formation-driven outflowsand their kinematics. The latter are important when considering under whichcircumstances matter energized by a starburst will reach the so-called ‘blowout’ condition, i.e. supersede the threshold energy starting at which local energy injection into the ISM can drive an outflow first into the halo (where metal re-distribution might be very efficient) and eventually out into intergalactic space. I will discuss here a few of these quantities, how we canmeasure them better than in the past, and in which way some of our observing techniques need to be improved in order toobtain better constraints from the data.     

Soft X—ray Properties of Ultraluminous IRAS Galaxies

We report on the results of cross-correlation of a sample of 903 Utraluminous IRAS galaxies (ULIRGs) with the ROSAT-All Sky Survey Bright Source Catalogue and the ROSAT archived pointing observations. The sample of ULIRGs has been compiled from the recently released PSCz redshift survey. In total,35 ULIRGs are securely detected by the ROSAT All-Sky Survey and pointing observations, five of which are blazars. The statistical properties of these sources in the soft X-ray band are determined and compared with their properties on other wavebands. We find that the ratio of the soft X-ray to the far-infrared flux spans about five orders of magnitude and reaches values of about unity. This ratio is a good indicator of the main energy source of ULIRGs. Those with soft X-ray to far-infrared flux exceeding 0.01 are probably powered by accretion onto central supermassive black holes while those with ratios smaller than 0.001 are probably powered by starbursts or other heating processes, or are Compton thick sources. Some ULIRGs have energy contributions from both. This ratio is low for most ULIRGs and hyperluminous infrared galaxies, which explains their low detection rate by ROSAT and ASCA.We also find that some ULIRGs have a similar soft X-ray luminosity vs. temperature relation to that for groups of galaxies and elliptical galaxies,suggesting a common origin of these systems. Our study also reveals a tight correlation between the hardness ratio and the soft X-ray luminosity for Seyfert 1s/QSOs.     

Multiwavelength study of 230 IRAS FSC galaxies

A sample of 230 galaxies has been compiled based on their IRAS FSC fluxes to study their multiwavelength properties and carry out comparative analyses with other similar samples. To understand the nature of these galaxies, a comparison with a sample of bright ULIRGs having fluxes at 60 μm fν > 1 Jy and 14 optically faint IRAS FSC galaxies is presented. This comparison shows that galaxies found by IRAS are not always strong infrared sources and that the objects from the sample of ULIRGs represent a sample of extreme galaxies, which are very powerful infrared sources. We have confirmed the consistency between Star Formation Rates (SFRs) calibrated based on luminosities of the PAH feature at 7.7 μm and radio 1.4 GHz flux densities. We have estimated the extinction for our 230 objects using SFRs calibrated from the PAH feature compared to ultraviolet flux, which shows that only 1% of the ultraviolet continuum typically escapes extinction by dust within a starburst.     

The far-infrared signature of dust in high-latitude regions

We present ISOPHOT observations of eight interstellar regions in the 60–200 μm wavelength range. The regions belong to mostly quiescent high-latitude clouds and have optical extinction peaks from   A_V ∼1–6 mag  . From the 150- and 200-μm emission, we derived colour temperatures for the classical big grain component which show a clear trend of decreasing temperature with increasing 200-μm emission. The 200-μm emission per unit   A_V   , however, does not drop at lower temperatures. This fact can be interpreted in terms of an increased far-infrared (FIR) emissivity of the big grains. We developed a two-component model including warm dust with the temperature of the diffuse interstellar medium (ISM) of   T = 17.5 K  , and cold dust with   T = 13.5 K  and FIR emissivity increased by a factor of >4. A mixture of the two components can reproduce the observed colour variations and the ratios   I ₂₀₀/ A_V   and  τ₂₀₀/ A_V   . The relative abundance of small grains with respect to the big grains shows significant variations from region to region at low column densities. However, in lines of sight of higher column density, our data indicate the disappearance of small grains, perhaps a signature of adsorption/coagulation of dust. The larger size and porous structure could also explain the increased FIR emissivity. Our results from eight independent regions suggest that these grains might be ubiquitous in the galactic ISM.     

Dense Gas, Chemistry and Star Formation in Luminous Galaxies

The molecular phase of the ISM constitutes the main source of fuel for the activity in starburst and AGNs. The physical conditions and chemical constitution of the molecular gas will change with, and respond to, the evolution of the activity. This paper includes a short discussion of the ¹²CO/¹³CO 1–0 line intensity ratio as a diagnostic tool of the molecular gas properties of luminous galaxies – paired with examples of high-resolution studies of how the line ratio varies within galaxies. A possible connection between the OH megamasers and galaxies with unusually high ¹²CO/¹³CO 1–0 line intensity ratios are also briefly discussed.The relative intensities of the dense gas tracers HNC, HCN, HCO⁺ and CN are a result of both chemistry and starburst evolution. The discussion on the interpretation of HNC 1–0 emission includes the importance of ion-neutral chemistry in a luminous starburst region. Finally, simple cartoon ISM models and how they can be applied to LIRGs and ULIRGs, are presented.     

HST/WFPC2 imaging of the QDOT ultraluminous infrared galaxy sample

We present HST WFPC2 V -band imaging for 23 ultraluminous infrared galaxies (ULIRGs) taken from the QDOT redshift survey. The fraction of sources observed to be interacting is 87 per cent. Most of the merging systems show a number of compact 'knots', whose colour and brightness differ substantially from their immediate surroundings. Colour maps for nine of the objects show a non-uniform colour structure. Features include blue regions located towards the centres of merging systems which are likely to be areas of enhanced star formation, and compact red regions which are likely to be dust shrouded starbursts or active galactic nuclei. The host galaxies of the quasi-stellar objects (QSOs) in the sample were found to be either interacting systems or ellipticals. Our data show no evidence that ULIRGs are a simple transition stage between galaxy mergers and QSOs. We propose an alternative model for ULIRGs based on the morphologies in our sample and previous N -body simulations. Under this model ULIRGs as a class are much more diverse than a simple transition between galaxy merger and QSO. The evolution of IR power source and merger morphology in ULIRGs is driven solely by the local environment and the morphologies of the merger progenitors.     

Gas phase processes affecting galactic evolution

Gas processes affecting star formation are reviewed with an emphasis on gravitational and magnetic instabilities as a source of turbulence. Gravitational instabilities are pervasive in a multi-phase medium, even for sub-threshold column densities, suggesting that only an ISM with a pure-warm phase can stop star formation. The instabilities generate turbulence, and this turbulence influences the structure and timing of star formation through its effect on the gas distribution and density. The final trigger for star formation is usually direct compression by another star or cluster. The star formation rate is apparently independent of the detailed mechanisms for star formation, and determined primarily by the total mass of gas in a dense form. If the density distribution function is a log-normal, as suggested by turbulence simulations, then this dense gas mass can be calculated and the star formation rate determined from first principles. The results suggest that only 10^-4 of the ISM mass actively participates in the star formation process and that this fraction does so because its density is larger than 10⁵ cm^-3, at which point several key processes affecting dynamical equilibrium begin to break down. This revised version was published online in August 2006 with corrections to the Cover Date.     

Processes affecting the composition and structure of silicate grains in cometary nuclei

Nucleation is a non-equilibrium process: the products of this process are seldom the most thermodynamically stable condensates but are instead those which form fastest. It should therefore not be surprising that grains formed in a circumstellar outflow will undergo some degree of metamorphism if they are annealed or are exposed to a chemically active reagent. Metamorphism of refractory particles continues in the interstellar medium (ISM) where the driving forces are sputtering by cosmic ray particles, annealing by high energy photons and grain destruction in supernova generated shocks. Studies of the depletion of the elements from the gas phase of the interstellar medium tell us that if grain destruction occurs with high efficiency in the ISM, then there must be some mechanism by which grains can be formed in the ISM. Various workers have shown that refractory mantles could form on refractory cores by radiation processing of organic ices. A similar process may operate to produce refractory inorganic mantles on grain cores which survived the supernova shocks. Most grains in a cloud which collapses to form a star will be destroyed; many of the surviving grains will be severely processed. Grains in the outermost regions of the nebula may survive relatively unchanged by thermal processing or hydration. It is these grains which we hope to find in comets. However, only those grains encased in ice at low temperature can be considered pristine since a considerable degree of hydrous alteration might occur in a cometary regolith if the comet enters the inner solar system. Some discussion of the physical, chemical and isotopic properties of a refractory grain at each stage of its life cycle will be attempted based on the limited laboratory data available to date. Suggestions will be made concerning types of experimental data which are needed in order to better understand the processing history of cosmic dust.     

On the nature of high luminosity ULIRGs

We present the preliminary results for a sample of 21 high luminosity ULIRGs (L_IR > 10^12.3 L⊙), selected from the 1Jy sample. Based on R band imaging we have performed an analysis of their morphologies. The main result obtained is that about half of the host galaxies are represented by spheroids and in the other half an exponential disk has been detected or they show still trace of a merging process going on. Regarding the relation between morphology and nuclear activity it seems that more advanced mergers appear in the most active galaxies. Hence the data support the evolutionary scheme proposed by Sanders 15 years ago that ULIRGs can be the earliest precursors of optical QSOs. This revised version was published online in August 2006 with corrections to the Cover Date.     

Confirming a population of hot-dust dominated, star-forming, ultraluminous galaxies at high redshift

We identify eight   z > 1  radio sources undetected at 850 μm but robustly detected at 70 μm, confirming that they represent ultraluminous infrared galaxies (ULIRGs) with hotter dust temperatures  (〈 T _d〉= 52 ± 10 K)  than submillimetre galaxies (SMGs) at similar luminosities and redshifts. These galaxies share many properties with SMGs: ultraviolet spectra consistent with starbursts, high stellar masses and radio luminosities. We can attribute their radio emission to star formation since high-resolution Multi-Element Radio Linked Interferometer Network (MERLIN) radio maps show extended emission regions (with characteristic radii of 2–3 kpc), which are unlikely to be generated by active galactic nucleus (AGN) activity. These observations provide the first direct confirmation of hot, dusty ULIRGs which are missed by current submillimetre surveys. They have significant implications for future observations from the Herschel Space Observatory and Submillimetre Common-User Bolometer Array 2 (SCUBA2), which will select high-redshift luminous galaxies with less selection biases.     

The ultraluminous and hyperluminous infrared galaxies in the Sloan Digital Sky Survey, 2dF Galaxy Redshift Survey and 6dF Galaxy Survey

We present a result of cross-correlating the Infrared Astronomical Satellite Faint Source Catalogue with the spectroscopic catalogues of galaxies in the Fourth Data Release of the Sloan Digital Sky Survey, the Final Data Release of the 2dF Galaxy Redshift Survey (2dFGRS) and the Second Data Release of the 6dF Galaxy Survey. We have identified 324 ultraluminous infrared galaxies (ULIRGs) including 190 newly discovered ULIRGs, and two hyperluminous infrared galaxies. Adding these new ULIRGs, we increase the number of known ULIRGs by about 30 per cent. The reliability of the cross-correlation is estimated using the likelihood ratio method. The incompleteness of our sample introduced by the identification procedure in this study is estimated to be about 5 per cent. Our sample covers the redshift range of   z = 0.037–0.517  with a median redshift of     , which is larger than that     of the sample of previously known ULIRGs.     

Monsters and babies from the first / iras survey

Radio continuum emission at cm wavelengths is relatively little affectedby extinction. When combined with far-infrared (FIR) surveys thisprovides for a convenient and unbiased method to select (radio-loud)AGN and starbursts deeply embedded in gas and dust–rich galaxies. Suchradio-selected FIR samples are useful for detailed investigations ofthe complex relationships between (radio) galaxy and starburst activity, and to determine whether ULIRGs are powered by hidden quasars (monsters) or young stars (babies).We present the results of a large program to obtain identifications andspectra of radio-selected, optically faint IRAS/FSC objects using theFIRST/VLA 20 cm survey (Becker, White and Helfand 1995). These objects are all radio-`quiet' in the sense that their radiopower / FIR luminosities follow the well-known radio/FIR relationshipfor star forming galaxies.We compare these results to a previous study by our group of a sampleof radio-`loud' IRAS/FSC ULIRGs selected from the Texas 365 MHzsurvey (Douglas etal. 1996). Many of these objects alsoshow evidence for dominant, A-type stellar populations, as well as highionization lines usually associated with AGN. These radio-loud ULIRGshave properties intermediate between those of starbursts and quasars,suggesting a possible evolutionary connection.Deep Keck spectroscopic observations of three ULIRGs from these samplesare presented, including high signal-to-noise spectropolarimetry.The polarimetry observations failed to show evidence of a hidden quasar inpolarized (scattered) light in the two systems in which the stellar lightwas dominated by A-type stars. Although observations of a larger samplewould be needed to allow a general conclusion, our current data suggestthat a large fraction of ULIRGs may be powered by luminous starbursts,not by hidden, luminous AGN (quasars).     

Continuum Evn and Merlin Observations of Ultra Luminous Infrared Galaxies

Radio imaging of ULIR galaxies is ideal to explore the connection between the starburst and the AGN phenomenon since it is unaffected by dust obscuration, and provides the required high angular resolution to distinguish between an AGN and starburst emission. We have made combined 18 cm radio continuum, EVN and MERLIN observations of 13 ULIRGs that have the parsec and deci-parsec scale resolution necessary to distinguish between an AGN and supernovae remnants at the centres of these galaxies, and assess the contribution of each to the total energy distribution. Images of three galaxies are presented here.     

Dense molecular gas in a sample of LIRGs and ULIRGs: The low-redshift connection to the huge high-redshift starbursts and AGNs

The sample of nearby LIRGs and ULIRGs for which dense molecular gas tracers have been measured is building up, allowing for the study of the physical and chemical properties of the gas in the variety of objects in which the most intense star formation and/or AGN activity in the local universe is taking place. This characterisation is essential to understand the processes involved, discard others and help to interpret the powerful starbursts and AGNs at high redshift that are currently being discovered and that will routinely be mapped by ALMA. We have studied the properties of the dense molecular gas in a sample of 17 nearby LIRGs and ULIRGs through millimeter observations of several molecules (HCO⁺, HCN, CN, HNC and CS) that trace different physical and chemical conditions of the dense gas in these extreme objects. In this paper we present the results of our HCO⁺ and HCN observations. We conclude that the very large range of measured line luminosity ratios for these two molecules severely questions the use of a unique molecular tracer to derive the dense gas mass in these galaxies.     

Fragmentation and Structure Formation

I present an overview of the hierarchy of structures existing in the interstellar medium (ISM) and the possible mechanisms that cause the fragmentation of one level into the next, with the formation of stars as its last step. Within this framework, I then give an overview of the contributions to this session. Numerical work addresses, at the largest scales, the shaping and formation of structures in the ISM through turbulence driven by stellar energy injection, and the resulting star formation rate as a function of mean density. At the scales of molecular clouds, results comparing observational and numerical data on the density and velocity structure of turbulence-produced cores, as well as their mass spectra, are summarized, together with existing theories of core and star formation controlled by the turbulence. Observationally, an attempt to discriminate between the standard and turbulent models of star formation is presented, finding inconclusive results, but suggesting that both turbulence and the magnetic field are dynamically important in molecular clouds and their cores. Finally, various determinations of the magnetic field strength and geometry are also presented.     

Galactic winds and circulation of the interstellar medium in dwarf galaxies

We study, through 2D hydrodynamical simulations, the feedback of a starburst on the ISM of typical gas-rich dwarf galaxies. The main goal is to address the circulation of the ISM and metals following the starburst. We assume a single-phase rotating ISM in equilibrium in the galactic potential generated by a stellar disc and a spherical dark halo. The starburst is assumed to occur in a small volume in the centre of the galaxy, and it generates a mechanical power of 3.8×10³⁹ or 3.8×10⁴⁰ erg s⁻¹ for 30 Myr. We find, in accordance with previous investigations, that the galactic wind is not very effective in removing the ISM. The metal-rich stellar ejecta, however, can be efficiently expelled from the galaxy and dispersed in the intergalactic medium.
Moreover, we find that the central region of the galaxy is always replenished with cold and dense gas a few 100 million years after the starburst, achieving the requisite for a new star formation event in ≈0.5–1 Gyr. The hydrodynamical evolution of galactic winds is thus consistent with the episodic star formation regime suggested by many chemical evolution studies.
We also discuss the X-ray emission of these galaxies and find that the observable (emission-averaged) abundance of the hot gas underestimates the real one if thermal conduction is effective. This could explain the very low hot-gas metallicities estimated in starburst galaxies.     

The injection of energy into the interstellar medium by stars (invited review)

Stars inject energy into the interstellar medium (ISM) by radiation, stellar winds, and supernova explosions. This energy injection causes the ISM to be inhomogeneous, which in turn alters the manner in which the energy is transferred through the ISM. A significant fraction of the energy is injected by massive stars, which formHii regions in the ISM. The structure and evolution ofHii regions in a cloudy medium deffers significantly from that in a homogeneous one. The strong stellar winds produced by massive stars form bubbles in the ISM, and the structure of these bubbles is often dominated by the structure of theHii region in which they are embedded. Finally, when the star explodes as a supernova, the evolution and appearance of the resulting remnant is determined by the structure of the bubble andHii region formed by the star during its lifetime.Paper presented at the IAU Third Asian-Pacific Regional Meeting, held in Kyoto, Japan, between 30 September–6 October, 1984.     

Programs for laser-AO assisted integral-field spectrometers on ionized flows

An AO-assisted integral-field spectrograph is becoming the most efficient tool with which to explore ionized gas outflows. It maps faint spectral lines that diagnose cloud dust content, gas pressure, excitation mechanism, and chemical abundances. Coupled with recent improvements in photoionization models, the total mass hence flow energetics can be estimated. Establishing a consistent dynamical framework requires linking multi-frequency datasets to track the energy flow through its optimal-contrast emission in the various ISM phases. I show HST results on AGN, starburst nuclei, and Galactic Herbig-Haro Objects that need complementary 3D spectra at comparable spatial resolution to come soon from laser-guided AO + integral-field spectrographs at the William-Herschel and SOAR telescopes.     

A quantitative determination of the AGN content in local ULIRGs through L-band spectroscopy

We present a quantitative estimate of the relative active galactic nucleus (AGN)/starburst content in a sample of 59 nearby  ( z < 0.15)  infrared bright ultraluminous infrared galaxies (ULIRGs) taken from the 1-Jy sample, based on infrared L -band (3–4 μm) spectra. By using diagnostic diagrams and a simple deconvolution model, we show that at least 60 per cent of local ULIRGs contain an active nucleus, but the AGN contribution to the bolometric luminosity is relevant only in  ∼15–20  per cent of the sources. Overall, ULIRGs appear to be powered by the starburst process, responsible for >85 per cent of the observed infrared luminosity. The subsample of sources optically classified as low-ionization nuclear emission-line regions (LINERs; 31 objects) shows a similar AGN/starburst distribution as the whole sample, indicating a composite nature for this class of objects. We also show that a few ULIRGs, optically classified as starbursts, have L -band spectral features suggesting the presence of a buried AGN.     

Physical conditions in radio jets and re-acceleration of relativistic electrons by MHD turbulence

Based on the observed features and physical conditions in the radio jets of several low-luminosity radio galaxies, I discuss the re-acceleration of relativistic electrons. On assuming a Fermi type acceleration, an acceleration coefficient of ～10^?15 s^?1 was obtained, which can well explain the radio brightness distribution in the jets and their spectrum. I further discuss the possibility of MHD turbulence providing the acceleration, and find that the turbulence energy spectral index must be restricted to the very narrow range 1.6–1.7.