Follow-up near-infrared spectroscopy of ultraluminous infrared galaxies observed by ISO

We present low-resolution mid-infrared (MIR) spectra of 16 ultraluminous infrared galaxies (ULIRGs) obtained with the circular variable filter (CVF) spectroscopy mode of ISOCAM on board the Infrared Space Observatory (ISO). Our sample completes previous ISO spectroscopy of ultra- and hyperluminous infrared galaxies toward higher luminosities. The combined samples cover an infrared luminosity range of ~1012-1013.1 L☉. To discriminate active galactic nucleus (AGN) and starburst activity, we use the AGN-related MIR continuum and the starburst-related 6.2, 7.7, 8.6, and 11.3 μm MIR emission bands attributed to aromatic carbonaceous material. For about half of the high-luminosity ULIRGs studied here, strong aromatic emission bands suggest starburst dominance. Other spectra are dominated by a strong AGN-related continuum with weak superposed emission features of uncertain nature. Our sample contains one unusual example, IRAS F00183-7111, of an AGN that is highly obscured even in the MIR. An improved method to characterize quantitatively the relative contribution of star formation and AGN activity to the MIR emission of ULIRGs is presented. The ULIRG spectra are fitted by a superposition of a starburst and an AGN spectrum, both of which may be obscured at different levels. Models in which starburst and AGN obscuration differ are significantly more successful than models with a single extinction. Previous results based on a simpler line-to-continuum measure of aromatic emission strength are confirmed, further supporting the robustness of the aromatic emission feature as a diagnostic of ULIRG power sources. As dominant sources of the bolometric luminosity, starbursts prevail at the lower end and AGNs at the higher end of this range. The transition between mostly starburst and mostly AGN powered occurs at ~1012.4-1012.5 L☉, and individual luminous starbursts are found up to ~1012.65 L☉.

We present a low-resolution mid-infrared spectroscopic survey of an unbiased sample of 62 ultraluminous infrared galaxies (ULIRGs) (LIR > 1012L⊙, z ≤ 0.3) using ISOPHOT-S on board the Infrared Space Observatory (ISO). For comparison, we also present ISOPHOT-S spectra for 23 active galactic nuclei (AGNs) and 15 starburst and normal galaxies. The line-to-continuum ratio of the 7.7 μm polycyclic aromatic hydrocarbon (PAH) emission feature is used as a discriminator between starburst and AGN activity in ULIRGs. We find that the majority of ULIRGs are predominantly powered by starbursts. The ratio of PAH over infrared luminosities, LPAH/LIR, for starburst-dominated ULIRGs is very similar to the ratio found for template starbursts. The shapes of the PAH features are sometimes unusual. Extinction has a noticeable effect on the PAH spectra of ULIRG starbursts. We have obtained high-resolution near-infrared imaging for the southern ISOPHOT-S ULIRGs in order to investigate their evolution stage. The majority (68%) of the ULIRGs imaged are double systems, and all show distorted morphologies. Of the 23 double-nuclei systems, 17 appear at linear separations between 4 and 14 kpc, with a mean separation of 6.5 kpc. Using the separations measured from our new near-infrared imaging as well as from the literature, we have examined whether ULIRGs that are advanced mergers are more AGN-like. We have found no such evidence, contrary to what is postulated by the classical evolutionary scenario. No correlation is found between the stage of merger in ULIRGs and their infrared luminosity. In fact, we find that systems in the early stages of merging may well put out maximum luminosity. We also find that the total mass of interstellar gas, as estimated from the CO (1 → 0) luminosity, does not decrease with decreasing merger separation. When both an AGN and a starburst occur concurrently in ULIRGs, we find that the starburst dominates the luminosity output. We propose that the available gas reservoir and the individual structure of the interacting galaxies plays a major role in the evolution of the system.

New tools from Infrared Space Observatory (ISO) mid-infrared spectroscopy have recently become available to determine the power sources of dust-obscured ultraluminous infrared galaxies (ULIRGs). We compare ISO classifications—starburst or active galactic nucleus (AGN)—with classifications from optical spectroscopy and with optical/near-infrared searches for hidden broad-line regions. The agreement between mid-infrared and optical classification is excellent if optical LINER spectra are assigned to the starburst group. The starburst nature of ULIRG LINERs strongly supports the suggestion that LINER spectra in infrared-selected galaxies, rather than being an expression of the AGN phenomenon, are due to shocks that are probably related to galactic superwinds. Differences between ISO and optical classification provide clues on the evolution of ULIRGs and on the configuration of obscuring dust. We find few ISO AGNs with optical H II or LINER identification, suggesting that highly obscured AGNs exist but are not typical for the ULIRG phenomenon in general. Rather, our results indicate that strong AGN activity, once triggered, quickly breaks the obscuring screen at least in certain directions, thus becoming detectable over a wide wavelength range.

This thesis explores the molecular gas (the raw material for star formation), especially the dense molecular gas content of luminous infrared galaxies (LIGs) and ''normal'' galaxies using millimeter line observations. Most LIGs are closely interacting/merging galaxies and ''normal'' spiral galaxies are believed to be the building blocks of LIGs. We here study and compare the distributions and masses of the total molecular gas and the dense molecular gas, traced by CO and HCN emission respectively, in LIGs, starburst galaxies and ''normal'' spiral galaxies. The molecular gas properties are then compared with the stages of galaxy- galaxy interaction and the far-IR luminosity to understand the star formation process and evolution of LIGs. We first study the dense molecular gas extent and distribution in ~ 10 nearby ''normal'' galaxies by mapping HCN emission (complementary with CO) at least along the major axes. We present the first detailed observational evidence that HCN emission in galaxies, \\ie, the dense molecular gas, is not confined to the inner ~ 1 kpc nuclear region, although the highest concentrations of dense molecular gas are in the center. A significant fraction of dense molecular gas is distributed in the inner disks of galaxies outside the nuclear or inner ring starburst regions, and can be detected to radii as large as a few kpc, perhaps to diameters of ~ D25/4. Then we have further surveyed HCN emission in more than ~ 40 relatively distant IR/CO bright ''normal'' galaxies and LIGs. Therefore, a statistically significant sample of HCN data in galaxies is established. We find that LIGs, especially ultraluminous ones, contain tremendous amount of dense molecular gas fueling the starbursts. Although LIGs are rich in molecular gas, some ''normal'' galaxies could have as much molecular gas as LIGs. However, the dense molecular gas content of even gas-rich ''normal'' galaxies is much less than that of LIGs of comparable molecular gas content. We show that HCN emission is better correlated with IR emission than that of CO. The total dense molecular content, the ratio of HCN/CO luminosity and the distribution of the surface brightness ratio IHCN/ICO are important high mass star formation indicators. We also confirm that the star formation efficiency indicated by LIR/LCO depends on the fraction of dense molecular gas (LHCN/LCO) and that the LIR/LHCN ratio is similar in all galaxies, ultraluminous or not, hot or cold in dust temperature, illustrating the starburst nature of ultraluminous IR galaxies. A second goal of this thesis is to study LIGs in the intermediate merging process, to determine the relationship between the various IR/CO properties and galaxy-galaxy interactions. We find a correlation between the CO luminosity and the projected separation of merger nuclei in a sample of more than ~ 50 LIG mergers, which suggests that the molecular content is decreasing as merging advances. Although the correlation is weak for the truly ultraluminous (LIR>1012 ls ) IR mergers, which could simply be due to the incompleteness of the sample since ultraluminous mergers are at great distances, the correlation is better established with less luminous LIG mergers, more close to a volume-limited statistically complete sample. The correlation slope for this nearby almost complete sample is the same as that of the large, heterogeneous sample of 50 mergers. We conducted new CO observation in ~ 20 LIG mergers to provide the CO data for this statistically complete sample of nearby LIG mergers. This correlation seems to have important constraints on both the merger- induced star formation models and the evolution scenario of LIGs. We have also conducted high resolution interferometry CO imaging in two spectacular LIGs, Arp 118 and Arp 119. We detected strong CO emission from rings/tails more than 10 kpc away from the nuclei, and most molecular gas is extranuclear. Detailed study of the distribution and kinematics of the molecular gas can provide the dynamical clues to the origin and evolution of these LIGs. Finally, we study the environment and spatial distribution of LIGs. We found that LIGs have stronger clustering than that of \\IRAS galaxies, yet most of them are not inside the groups of galaxies. Using various previous studies along with our own present investigations, we suggest that some LIGs are remnants of the merged groups of gas-rich galaxies.

We present sub-mm photometry for 11 Hyperluminous Infrared Galaxies (HLIRGs) and use radiative transfer models for starbursts and AGN to investigate the IR emission. In all sources both a starburst and AGN are required to explain the IR emission. The mean starburst fraction is 35%, with a range spanning 80% starburst dominated to 80% AGN dominated. In all cases the starburst dominates at rest-frame wavelengths >50 microns, with star formation rates >500 solar masses per year. The trend of increasing AGN fraction with increasing IR luminosity seen in IRAS galaxies peaks in HLIRGs, and is not higher than the fraction seen in bright ULIRGs. The AGN and starburst luminosities correlate, suggesting that a common physical factor, plausibly the dust masses, governs their luminosities. Our results suggest that the HLIRG population is comprised both of ULIRG-like galaxy mergers, and of young galaxies going through their maximal star formation periods whilst harbouring an AGN. The coeval AGN and starburst activity in our sources implies that starburst and AGN activity, and the peak starburst and AGN luminosities, can be coeval in active galaxies generally. When extrapolated to high-z our sources have comparable sub-mm fluxes to sub-mm survey sources. At least some sub-mm survey sources are therefore likely to be comprised of similar galaxy populations to those found in the HLIRG population. It is also plausible from these results that high-z sub-mm sources harbour heavily obscured AGN. The differences in X-ray and sub-mm properties between HLIRGs at z~1 and sub-mm sources at z~3 implies evolution between the two epochs. Either the mean AGN obscuration level is greater at z~3 than at z~1, or the fraction of IR-luminous sources at z~3 that contain AGN is smaller than that at z~1.

We report measurements of the [C II] 157.74 μm fine-structure line in a sample of seven ultraluminous infrared galaxies (ULIGs) (LIR > 1012 L☉) with the Long Wavelength Spectrometer on the Infrared Space Observatory. The [C II] line is an important coolant in galaxies and arises in interstellar gas exposed to far-ultraviolet photons (hν≥11.26 eV); in ULIGs, this radiation stems from the bursts of star formation and/or from the active galactic nuclei that power the tremendous infrared luminosity. The [C II] 158 μm line is detected in four of the seven ULIGs; the absolute line flux (about a few times 10-20 W cm-2) represents some of the faintest extragalactic[C II] emission yet observed. Relative to the far-infrared continuum, the [C II] flux from the observed ULIGs is ~10% of that seen from nearby normal and starburst galaxies. We discuss possible causes for the [C II] deficit, namely (1) self-absorbed or optically thick [C II] emission, (2) saturation of the [C II] emission in photodissociated gas with high gas density n (3 × 103 cm-3) or with a high ratio of incident UV flux G0 to n (G0/n 10 cm3), or (3) the presence of a soft ultraviolet radiation field caused, for example, by a stellar population deficient in massive main-sequence stars. As nearby examples of colliding galaxies, ULIGs may resemble high-redshift protogalaxies in both morphology and spectral behavior. If true, the suggested [C II] deficit in ULIGs poses limitations on the detection rate of high-z sources and on the usefulness of [C II] as an eventual tracer of protogalaxies.

We present upper limits on the 850 micron and 450 micron fluxes of the warm hyperluminous (bolometric luminosity L_bol > 10^13 L_sun galaxies IRAS P09104+4109 (z=0.442) and IRAS F15307+3252 (z=0.926), derived from measurements using the SCUBA bolometer array on the James Clerk Maxwell Telescope. Hot luminous infrared sources like these are thought to differ from more normal cold ultraluminous infrared (L_bol > 10^12 L_sun) galaxies in that they derive most of their bolometric luminosities from dusty AGNs as opposed to starbursts. Such hot, dusty AGNs at high redshift are thought to be responsible for much of the mass accretion of the Universe that is in turn responsible for the formation of the supermassive black holes seen in the centres of local galaxies. The galaxy IRAS P09104+4109 is also unusual in that it is a cD galaxy in the center of a substantial cooling-flow cluster, not an isolated interacting galaxy like most ultraluminous infrared galaxies. Previously it was known to have large amounts of hot (T > 50 K) dust from IRAS observations. We now show that the contribution of cold dust to the bolometric luminosity is less than 3 per cent. Most ultraluminous infrared galaxies possess large amounts of cold dust, and it is now known that some cooling flow cluster cD galaxies do as well. Yet this object, which is an extreme example of both, does not have enough cold gas to contribute significantly to the bolometric luminosity. We outline physical reasons why this could have happened. We then provide a discussion of stategies for finding hot dusty AGNs, given the limitations on submillimetre surveys implied by this work.

We review investigations with SWS, ISOPHOT-S, and ISOCAM-CVF of several samples of ultra-luminous infrared galaxies (ULIRGs) drawn from IRAS catalogs. The mid-IR data indicate that the majority of the local ULIRG population is predominantly powered by star formation but that AGN activity plays a signi.cant role in many sources and dominates the most luminous ones. AGN and starburst activity is present in all phases of the ULIRG/merger evolution. There is no obvious trend for AGNs to be more predominant, or the molecular content to be lower in the late ULIRG phases. Star formation activity in ULIRGs likely occurs in short (107 years) phases terminated by the disruptive effects of superwinds and supernovae. Highly obscured AGNs dominating the total energy output may be hidden to the mid-infrared observations in a few cases but are not characteristic of the average ULIRG.

The massive cluster of galaxies Abell 2219 (z = 0.228) with two spectacular gravitational lensing arcs was observed at 14.3 μm (hereafter 15 μm) with the Infrared Space Observatory and results were published by Barvainis et al. ([CITE]). These observations have been reanalyzed using a method specifically designed for the detection of faint sources that had been applied to other clusters. Five new sources were detected and the resulting cumulative total of ten sources all have optical counterparts. The mid-infrared sources are identified with three cluster members, three foreground galaxies, an Extremely Red Object, a star and two galaxies of unknown redshift. The spectral energy distributions (SEDs) of the galaxies are fit with models from a selection, using the program GRASIL. Best-fits are obtained, in general, with models of galaxies with ongoing star formation. Infrared luminosities and star formation rates are obtained for six sources: the cluster members and the foreground galaxies. For the three cluster members the infrared luminosities derived from the model SEDs are between ~5.7 and 1.4 , corresponding to infrared star formation rates between 10 and 24 . The two cluster galaxies that have optical classifications are in the Butcher-Oemler region of the color–magnitude diagramme. The three foreground galaxies have infrared luminosities between 1.5 and 9.4 yielding infrared star formation rates between 3 and 16 . Two of the foreground galaxies are located in two foreground galaxy enhancements (Boschin et al. [CITE]). Including Abell 2219, six distant clusters of galaxies have been mapped with ISOCAM and luminous infrared galaxies (LIRGs) have been found in three of them. The presence of LIRGs in Abell 2219 strengthens the association between luminous infrared galaxies in clusters and recent or ongoing cluster merger activity.

We present initial results from a Chandra survey of a complete sample of the eight nearest (z ≤ 0.04) ultraluminous IR galaxies (ULIRGs) and also include the IR-luminous galaxy NGC 6240 for comparison. In this paper we use the hard X-rays (2-8 keV) to search for the possible presence of an obscured active galactic nucleus (AGN). In every case, a hard X-ray source is detected in the nuclear region. If we divide the sample according to the optical/IR spectroscopic classification (starburst vs. AGN), we find that the five ULIRGs have hard X-ray luminosities about an order of magnitude smaller than the three AGN ULIRGs. NGC 6240 has an anomalously high hard X-ray luminosity compared to the starburst ULIRGs. The Fe-Kα line is convincingly detected in only two ULIRGs. The weakness of the Fe-K emission in these ULIRGs generally suggests that the hard X-ray spectrum is not dominated by reflection from high-NH neutral material. The hard X-ray continuum flux ranges from a few × 10-3 to a few × 10-5 of the far-IR flux, similar to values in pure starbursts and several orders of magnitude smaller than in Compton-thin AGNs. The upper limits on the ratio of the Fe-Kα-to-far-IR flux are below the values measured in Compton-thick type 2 Seyfert galaxies. While very large column densities of molecular gas are observed in the nuclei of these galaxies, we find no evidence that the observed X-ray sources are obscured by Compton-thick material. Thus, our new hard X-ray data do not provide direct evidence that powerful buried quasars dominate the overall energetics of most ULIRGs.

We present the first results of a high-resolution Karl G. Jansky Very Large Array imaging survey of luminous and ultra-luminous infrared galaxies (U/LIRGs) in the Great Observatories All-sky LIRG Survey. From the full sample of 68 galaxies, we have selected 25 luminous infrared galaxies (LIRGs) that show resolved extended emission at sufficient sensitivity to image individual regions of star formation activity beyond the nucleus. With wideband radio continuum observations, which sample the frequency range from 3 to 33 GHz, we have made extinction-free measurements of the luminosities and spectral indicies for a total of 48 individual star-forming regions identified as having deprojected galactocentric radii (r G ) that lie outside the 13.2 μm core of the galaxy. The median 3–33 GHz spectral index and 33 GHz thermal fraction measured for these “extranuclear” regions is −0.51 ± 0.13 and 65% ± 11%, respectively. These values are consistent with measurements made on matched spatial scales in normal star-forming galaxies, and suggests that these regions are more heavily dominated by thermal free–free emission relative to the centers of local U/LIRGs. Further, we find that the median star formation rate derived for these regions is ∼1 M ⊙ yr−1, and when we place them on the sub-galactic star-forming main sequence of galaxies (SFMS), we find they are offset from their host galaxies’ globally averaged specific star formation rates. We conclude that while nuclear starburst activity drives LIRGs above the SFMS, extranuclear star formation still proceeds in a more extreme fashion relative to what is seen in local spiral galaxies.

We present radio observations of ultraluminous infrared galaxies (ULIRGs) using the Giant Metrewave Radio Telescope (GMRT) and combine them with archival multifrequency observations to understand whether ULIRGs are the progenitors of the powerful radio loud galaxies in the local Universe. ULIRGs are characterized by large infrared luminosities (LIR > 1012 L⊙), large dust masses (∼108 M⊙), and vigorous star formation (star formation rates ∼10–100 M⊙ yr−1). Studies show that they represent the end stages of mergers of gas-rich spiral galaxies. Their luminosity can be due to both starburst activity and active galactic nuclei (AGNs). We study a sample of 13 ULIRGs that have optically identified AGN characteristics with 1.28 GHz GMRT observations. Our aim is to resolve any core-jet structures or nuclear extensions and hence examine whether the ULIRGs are evolving into radio loud ellipticals. Our deep, low frequency observations show marginal extension for only one source. However, the integrated radio spectra of 9 ULIRGs show characteristics that are similar to that of GPS/CSS/CSO/young radio sources. The estimated spectral ages are 0.4–20 Myr and indicate that they are young radio sources and possible progenitors of radio galaxies. Hence, we conclude that although most ULIRGs do not show kpc scale extended radio emission associated with nuclear activity, their radio spectral energy distributions do show signatures of young radio galaxies.

We have established a model to systematically estimate the contribution of the mid-IR emission features between 3 and 11.6 μm to the IRAS in-band fluxes, using the results of Infrared Space Observatory PHT Spectrophotometer observation of 16 galaxies by Lu et al. in 1997. The model is used to estimate more properly the k-corrections for calculating the rest frame 12 and 25 μm fluxes and luminosities of IRAS galaxies. We have studied the 12-25 μm color-luminosity relation for a sample of galaxies selected at 25 μm. The color is found to correlate well with the 25 μm luminosity, the mid-IR luminosity, and the ratio of far-infrared and blue luminosities. The relations with the mid-IR luminosities are more sensitive to different populations of galaxies, while a single relation of the 12-25 μm color versus the ratio of the far-infrared and blue luminosities applies equally well to these different populations. The luminous and ultraluminous infrared galaxies have redder 12-25 μm colors than those of the quasars. These relations provide powerful tools to differentiate different populations of galaxies. The local luminosity function at 12 μm provides the basis for interpreting the results of deep mid-IR surveys planned or in progress with Infrared Space Observatory, Wide-Field Infrared Explorer, and Space Infrared Telescope Facility. We have selected a sample of 668 galaxies from the IRAS Faint Source Survey flux density limited at 200 mJy at 12 μm. A 12 μm local luminosity function is derived and, for the first time in the literature, effects of density variation in the local universe are considered and corrected in the calculation of the 12 μm luminosity function. It is also found that the 12 μm selected sample is dominated by quasars and active galaxies, which therefore strongly affect the 12 μm luminosity function at high luminosities. The ultraluminous infrared galaxies are relatively rare at 12 μm compared to a 25 μm sample.

We present 7–180 μm photometry of a sample of hyperluminous infrared galaxies (HyLIGs) obtained with the photometer and camera mounted on the Infrared Space Observatory. We have used radiative transfer models of obscured starbursts and dusty torii to model their spectral energy distributions (SEDs). We find that IRAS F00235+1024, IRAS F14218+3845 and IRAS F15307+3252 require a combination of starburst and active galactic nuclei (AGN) components to explain their mid- to far-infrared (FIR) emission, while for TXS 0052+471 a dust torus AGN model alone is sufficient. For IRAS F00235+1024 and IRAS F14218+3845 the starburst component is the predominant contributor, whereas for IRAS F15307+3252 the dust torus component dominates. The implied star formation rates (SFRs) for these three sources estimated from their infrared luminosities are yr−1h50−2 and are amongst the highest SFRs estimated to date. We also demonstrate that the well-known radio–FIR correlation extends into both higher radio and infrared power than previously investigated. The relation for HyLIGs has a mean q value of 1.94. The results of this study imply that better sampling of the infrared spectral energy distributions of HyLIGs may reveal that both AGN and starburst components are required to explain all the emission from the near-infrared to the submillimetre.

The Infrared Space Observatory (ISO) has for the first time made possible mid- and far-infrared spectroscopic studies of nearby and moderate redshift, dusty ultraluminous infrared galaxies (ULIRGs). We discuss recent SWS, LWS, ISOPHOT-S and ISOCAM (CVF) results on ULIRGs, addressing the following topics: What powers ULIRGs? What is the AGN-starburst connection? How do ULIRGs evolve? What is the connection of the local ULIRG population to the recently discovered population of mid-IR to submillimeter sources?