arXiv subjects

B. T. Soifer

Publications and source records attributed to B. T. Soifer.

At least 37 records · Page 2Linked to original sources

arXiv2008

The Origin of the 24-micron Excess in Red Galaxies

Observations with the Spitzer Space Telescope have revealed a population of red-sequence galaxies with a significant excess in their 24-micron emission compared to what is expected from an old stellar population. We identify 900 red galaxies with 0.15 0.3 mJy). We determine the prevalence of AGN and star-formation activity in all the AGES galaxies using optical line diagnostics and mid-IR color-color criteria. Using the IRAC color-color diagram from the IRAC Shallow Survey, we find that 64% of the 24-micron excess red galaxies are likely to have strong PAH emission features in the 8-micron IRAC band. This fraction is significantly larger than the 5% of red galaxies with f24<0.3 mJy that are estimated to have strong PAH emission, suggesting that the infrared emission is largely due to star-formation processes. Only 15% of the 24-micron excess red galaxies have optical line diagnostics characteristic of star-formation (64% are classified as AGN and 21% are unclassifiable). The difference between the optical and infrared results suggest that both AGN and star-formation activity is occurring simultaneously in many of the 24-micron excess red galaxies. These results should serve as a warning to studies that exclusively use optical line diagnostics to determine the dominant emission mechanism in the infrared and other bands. We find that ~40% of the 24-micron excess red galaxies are edge-on spiral galaxies with high optical extinctions. The remaining sources are likely to be red galaxies whose 24-micron emission comes from a combination of obscured AGN and star-formation activity.

astro-ph↗

arXiv2008

Morphologies of High Redshift, Dust Obscured Galaxies from Keck Laser Guide Star Adaptive Optics

Spitzer MIPS images in the Bootes field of the NOAO Deep Wide-Field Survey have revealed a class of extremely dust obscured galaxy (DOG) at z~2. The DOGs are defined by very red optical to mid-IR (observed-frame) colors, R - [24 um] > 14 mag, i.e. f_v (24 um) / f_v (R) > 1000. They are Ultra-Luminous Infrared Galaxies with L_8-1000 um > 10^12 -10^14 L_sun, but typically have very faint optical (rest-frame UV) fluxes. We imaged three DOGs with the Keck Laser Guide Star Adaptive Optics (LGSAO) system, obtaining ~0.06'' resolution in the K'-band. One system was dominated by a point source, while the other two were clearly resolved. Of the resolved sources, one can be modeled as a exponential disk system. The other is consistent with a de Vaucouleurs profile typical of elliptical galaxies. The non-parametric measures of their concentration and asymmetry, show the DOGs to be both compact and smooth. The AO images rule out double nuclei with separations of greater than 0.1'' (< 1 kpc at z=2), making it unlikely that ongoing major mergers (mass ratios of 1/3 and greater) are triggering the high IR luminosities. By contrast, high resolution images of z~2 SCUBA sources tend to show multiple components and a higher degree of asymmetry. We compare near-IR morphologies of the DOGs with a set of z=1 luminous infrared galaxies (LIRGs; L_IR ~ 10^11 L_sun) imaged with Keck LGSAO by the Center for Adaptive Optics Treasury Survey. The DOGs in our sample have significantly smaller effective radii, ~1/4 the size of the z=1 LIRGs, and tend towards higher concentrations. The small sizes and high concentrations may help explain the globally obscured rest-frame blue-to-UV emission of the DOGs.

astro-ph↗

arXiv2008

Redshift Distribution of Extragalactic 24 micron Sources

We present the redshift distribution of a complete, unbiased sample of 24 micron sources down to fnu(24 micron) = 300 uJy (5-sigma). The sample consists of 591 sources detected in the Bootes field of the NOAO Deep Wide-Field Survey. We have obtained optical spectroscopic redshifts for 421 sources (71%). These have a redshift distribution peaking at z~0.3, with a possible additional peak at z~0.9, and objects detected out to z=4.5. The spectra of the remaining 170 (29%) exhibit no strong emission lines from which to determine a redshift. We develop an algorithm to estimate the redshift distribution of these sources, based on the assumption that they have emission lines but that these lines are not observable due to the limited wavelength coverage of our spectroscopic observations. The redshift distribution derived from all 591 sources exhibits an additional peak of extremely luminous (L(8-1000 micron) > 3 x 10^{12} Lsun) objects at z~2, consisting primarily of sources without observable emission lines. We use optical line diagnostics and IRAC colors to estimate that 55% of the sources within this peak are AGN-dominated. We compare our results to published models of the evolution of infrared luminous galaxies. The models which best reproduce our observations predict a large population of star-formation dominated ULIRGs at z > 1.5 rather than the AGN-dominated sources we observe.

astro-ph↗

arXiv2008

Spitzer Mid-Infrared Spectroscopy of Distant X-ray Luminous AGN

We present mid-infrared spectroscopy of a sample of 16 optically faint infrared luminous galaxies obtained with the Infrared Spectrograph (IRS) on the Spitzer Space Telescope. These sources were jointly selected from Spitzer and Chandra imaging surveys in the NDWFS Bootes field and were selected from their bright X-ray fluxes to host luminous AGN. None of the spectra show significant emission from polycyclic aromatic hydrocarbons (PAHs; 6.2um equivalent widths <0.2um), consistent with their infrared emission being dominated by AGN. Nine of the X-ray sources show 9.7um silicate absorption features. Their redshifts are in the range 0.9<z<2.6, implying infrared luminosities of log(L{IR})=12.5-13.6 solar luminosities. The average silicate absorption strength is not as strong as that of previously targeted optically faint infrared luminous galaxies with similar mid-infrared luminosities implying that the X-ray selection favors sources behind a smaller column of Si-rich dust than non-X-ray selection. Seven of the X-ray sources have featureless power-law mid-IR spectra. We argue that the featureless spectra likely result from the sources having weak or absent silicate and PAH features rather than the sources lying at higher redshifts where these features are shifted out of the IRS spectral window. We investigate whether there are any correlations between X-ray and infrared properties and find that sources with silicate absorption features tend to have fainter X-ray fluxes and harder X-ray spectra, indicating a weak relation between the amount of silicate absorption and column density of X-ray-absorbing gas.

astro-ph↗

arXiv2008

A Significant Population of Very Luminous Dust-Obscured Galaxies at Redshift z ~ 2

Observations with Spitzer Space Telescope have recently revealed a significant population of high-redshift z~2 dust-obscured galaxies (DOGs) with large mid-IR to UV luminosity ratios. These galaxies have been missed in traditional optical studies of the distant universe. We present a simple method for selecting this high-z population based solely on the ratio of the observed mid-IR 24um to optical R-band flux density. In the 8.6 sq.deg Bootes NDWFS Field, we uncover ~2,600 DOG candidates (= 0.089/sq.arcmin) with 24um flux densities F24>0.3mJy and (R-[24])>14 (i.e., F[24]/F[R] > 1000). These galaxies have no counterparts in the local universe, and become a larger fraction of the population at fainter F24, representing 13% of the sources at 0.3~mJy. DOGs exhibit evidence of both star-formation and AGN activity, with the brighter 24um sources being more AGN- dominated. We have measured spectroscopic redshifts for 86 DOGs, and find a broad z distribution centered at ~2.0. Their space density is 2.82E-5 per cubic Mpc, similar to that of bright sub-mm-selected galaxies at z~2. These redshifts imply very large luminosities LIR>~1E12-14 Lsun. DOGs contribute ~45-100% of the IR luminosity density contributed by all z~2 ULIRGs, suggesting that our simple selection criterion identifies the bulk of z~2 ULIRGs. DOGs may be the progenitors of ~4L* present-day galaxies seen undergoing a luminous,short- lived phase of bulge and black hole growth. They may represent a brief evolution phase between SMGs and less obscured quasars or galaxies. [Abridged]

astro-ph↗

arXiv2007

Lens Galaxy Properties of SBS1520+530: Insights from Keck Spectroscopy and AO Imaging

We report on an investigation of the SBS 1520+530 gravitational lens system and its environment using archival HST imaging, Keck spectroscopic data, and Keck adaptive-optics imaging. The AO imaging has allowed us to fix the lens galaxy properties with a high degree of precision when performing the lens modeling, and the data indicate that the lens has an elliptical morphology and perhaps a disk. The new spectroscopic data suggest that previous determinations of the lens redshift may be incorrect, and we report an updated, though inconclusive, value z_lens = 0.761. We have also spectroscopically confirmed the existence of several galaxy groups at approximately the redshift of the lens system. We create new models of the lens system that explicitly account for the environment of the lens, and we also include improved constraints on the lensing galaxy from our adaptive-optics imaging. Lens models created with these new data can be well-fit with a steeper than isothermal mass slope (alpha = 2.29, with the density proportional to r^-alpha) if H_0 is fixed at 72 km/s/Mpc; isothermal models require H_0 ~ 50 km/s/Mpc. The steepened profile may indicate that the lens is in a transient perturbed state caused by interactions with a nearby galaxy.

astro-ph↗

arXiv2007

Spitzer Mid-Infrared Spectroscopy of 70um-Selected Distant Luminous Infrared Galaxies

We present mid-infrared spectroscopy obtained with the Spitzer Space Telescope of a sample of 11 optically faint, infrared luminous galaxies selected from a Spitzer MIPS 70um imaging survey of the NDWFS Bootes field. These are the first Spitzer IRS spectra presented of distant 70um-selected sources. All the galaxies lie at redshifts 0.3<z<1.3 and have very large infrared luminosities of L_IR~ 0.1-17 x 10^12 solar luminosities. Seven of the galaxies exhibit strong emission features attributed to polycyclic aromatic hydrocarbons (PAHs). The average IRS spectrum of these sources is characteristic of classical starburst galaxies, but with much larger infrared luminosities. The PAH luminosities of L(7.7) ~ 0.4 - 7 x 10^11 solar luminosities imply star formation rates of ~ 40 - 720 solar masses per year. Four of the galaxies show deep 9.7um silicate absorption features and no significant PAH emission features (6.2um equivalent widths < 0.03um). The large infrared luminosities and low f70/f24 flux density ratios suggests that these sources have AGN as the dominant origin of their large mid-infrared luminosities, although deeply embedded but luminous starbursts cannot be ruled out. If the absorbed sources are AGN-dominated, a significant fraction of all far-infrared bright, optically faint sources may be dominated by AGN.

astro-ph↗

arXiv2007

PAH Emission from Ultraluminous Infrared Galaxies

We explore the relationships between the Polycyclic Aromatic Hydrocarbon (PAH) feature strengths, mid-infrared continuum luminosities, far-infrared spectral slopes, optical spectroscopic classifications, and silicate optical depths within a sample of 107 ULIRGs observed with the Infrared Spectrograph on the Spitzer Space Telescope. The detected 6.2 micron PAH equivalent widths (EQWs) in the sample span more than two orders of magnitude (0.006-0.8 micron), and ULIRGs with HII-like optical spectra or steep far-infrared spectral slopes (S_{25} / S_{60} < 0.2) typically have 6.2 micron PAH EQWs that are half that of lower-luminosity starbursts. A significant fraction (~40-60%) of HII-like, LINER-like, and cold ULIRGs have very weak PAH EQWs. Many of these ULIRGs also have large (tau_{9.7} > 2.3) silicate optical depths. The far-infrared spectral slope is strongly correlated with PAH EQW, but not with silicate optical depth. In addition, the PAH EQW decreases with increasing rest-frame 24 micron luminosity. We argue that this trend results primarily from dilution of the PAH EQW by continuum emission from dust heated by a compact central source, probably an AGN. High luminosity, high-redshift sources studied with Spitzer appear to have a much larger range in PAH EQW than seen in local ULIRGs, which is consistent with extremely luminous starburst systems being absent at low redshift, but present at early epochs.

astro-ph↗

arXiv2007

High resolution mid-infrared spectroscopy of ultraluminous infrared galaxies

(Abridged) We present R~600, 10-37um spectra of 53 ULIRGs at z<0.32, taken using the IRS on board Spitzer. All of the spectra show fine structure emission lines of Ne, O, S, Si and Ar, as well as molecular Hydrogen lines. Some ULIRGs also show emission lines of Cl, Fe, P, and atomic Hydrogen, and/or absorption features from C_2H_2, HCN, and OH. We employ diagnostics based on the fine-structure lines, as well as the EWs and luminosities of PAH features and the strength of the 9.7um silicate absorption feature (S_sil), to explore the power source behind the infrared emission in ULIRGs. We show that the IR emission from the majority of ULIRGs is powered mostly by star formation, with only ~20% of ULIRGs hosting an AGN with a comparable or greater IR luminosity than the starburst. The detection of the 14.32um [NeV] line in just under half the sample however implies that an AGN contributes significantly to the mid-IR flux in ~42% of ULIRGs. The emission line ratios, luminosities and PAH EWs are consistent with the starbursts and AGN in ULIRGs being more extincted, and for the starbursts more compact, versions of those in lower luminosity systems. The excitations and electron densities in the NLRs of ULIRGs appear comparable to those of lower luminosity starbursts, though there is evidence that the NLR gas in ULIRGs is more dense. We show that the combined luminosity of the 12.81um [NeII] and 15.56um [NeIII] lines correlates with both IR luminosity and the luminosity of the 6.2 micron and 11.2 micron PAH features in ULIRGs, and use this to derive a calibration between PAH luminosity and star formation rate. Finally, we show that ULIRGs with 0.8 < S_sil < 2.4 are likely to be powered mainly by star formation, but that ULIRGs with S_sil < 0.8, and possibly those with S_sil > 2.4, contain an IR-luminous AGN.

astro-ph↗

arXiv2007

High resolution imaging of the anomalous flux-ratio gravitational lens system CLASS B2045+265: Dark or luminous satellites?

The existence of flux-ratio anomalies between fold and cusp images in galaxy-scale strong-lens systems has led to an interpretation based on the presence of a high mass-fraction of cold-dark-matter (CDM) substructures around galaxies, as predicted by numerical N-body simulations. The flux-ratio anomaly is particularly evident in the radio-loud quadruple gravitational lens system CLASS B2045+265. In this paper, new high-resolution radio, optical, and infrared imaging of B2045+265 is presented which sheds more light on this anomaly and its possible causes. First, deep Very Long Baseline Array observations show very compact images, possibly with a hint of a jet, but with no evidence for differential scattering or scatter broadening. Second, optical and infrared observations with the Hubble Space Telescope and through Adaptive-Optics imaging with the W. M. Keck Telescope, show a previously undiscovered object -- interpreted as a (tidally disrupted) dwarf satellite based on its colours and slight extension -- between the main lens galaxy and the three anomalous flux-ratio images. Third, colour variations in the early-type lens galaxy indicate recent star-formation, possibly the result of secondary infall of gas-rich satellites. A population of such galaxies around the lens system could explain the previously discovered strong [O II] emission. However, spiral structure and/or normal star formation in the lens galaxy cannot be excluded. In light of these new data, we propose a lens model for the system, including the observed dwarf satellite, which reproduces all positional and flux-ratio constraints, without the need for additional CDM substructure. [abridged]

astro-ph↗

arXiv2007

Optical Line Diagnostics of z~2 Optically Faint ULIRGs in the Spitzer Bootes Survey

We present near-infrared spectroscopic observations for a sample of ten optically faint luminous infrared galaxies (R-[24]> 14) using Keck NIRSPEC and Gemini NIRI. The sample is selected from a 24 micron Spitzer MIPS imaging survey of the NDWFS Bootes field. We measure accurate redshifts in the range 1.3 1900 km/s) Halpha or Hbeta emission lines; the remaining three are type II AGN. Given their large mid-IR luminosities and faint optical magnitudes, we might expect these sources to be heavily extincted quasars, and therefore only visible as type II AGN. The visibility of broad lines in 70% of the sources suggests that it is unlikely that these AGN are being viewed through the mid-plane of a dusty torus. For four of the sources we constrain the Halpha/Hbeta Balmer decrement and estimate the extinction to the emission line region to be large for both type I and type II AGN, with A_Halpha > 2.4-5 mag. Since the narrow-line region is also extincted and the UV continuum emission from the host galaxies is extremely faint, this suggests that much of the obscuration is contributed by dust on large (~kpc) scales within the host galaxies. These sources may be examples of "host-obscured" AGN which could have space densities comparable or greater to that of optically luminous type I AGN with similar bolometric luminosities.

astro-ph↗

arXiv2006

The 1<z<5 Infrared Luminosity Function of Type I Quasars

We determine the rest-frame 8 micron luminosity function of type I quasars over the redshift range 1<z<5. Our sample consists of 292 24 micron sources brighter than 1 mJy selected from 7.17 square degrees of the Spitzer Space Telescope MIPS survey of the NOAO Deep Wide-Field Survey Bootes field. The AGN and Galaxy Evolution Survey (AGES) has measured redshifts for 270 of the R<21.7 sources and we estimate that the contamination of the remaining 22 sources by stars and galaxies is low. We are able to select quasars missed by ultra-violet excess quasar surveys, including reddened type I quasars and 2.2<z<3.0 quasars with optical colors similar to main sequence stars. We find reddened type I quasars comprise 20% of the type I quasar population. Nonetheless, the shape, normalization, and evolution of the rest-frame 8 micron luminosity function is comparable to that of quasars selected from optical surveys. The 8 micron luminosity function of type I quasars is well approximated by a power-law with index -2.75(+/-0.14). We directly measure the peak of the quasar space density to be at z=2.6(+/-0.3).

astro-ph↗

arXiv2006

Observations of Ultraluminous Infrared Galaxies with the Infrared Spectrograph on the Spitzer Space Telescope II: The IRAS Bright Galaxy Sample

We present spectra taken with the Infrared Spectrograph on Spitzer covering the 5-38 micron region of the ten Ultraluminous Infrared Galaxies (ULIRGs) found in the IRAS Bright Galaxy Sample. Among the BGS ULIRGs, we find a factor of 50 spread in the rest-frame mid to far-infrared spectral slope. The 9.7 micron silicate optical depths range from less than 0.4 more than 4.2, implying line of sight extinctions of A(V) ~ 8 - 78 mag. There is evidence for water ice and hydrocarbon absorption and C2H2 and HCN absorption features in four and possibly six of the 10 BGS ULIRGs, indicating shielded molecular clouds and a warm, dense ISM. We have detected [NeV] emission in three of the ten BGS ULIRGs, at flux levels of 5-18E-14 erg/cm^2/sec and [NeV] 14.3/[NeII] 12.8 line flux ratios of 0.12-0.85. The remaining BGS ULIRGs have limits on their [NeV]/[NeII] line flux ratios which range from less than 0.15 to less than 0.01. Among the BGS ULIRGs, the AGN fractions implied by either the [NeV]/[NeII] or [OIV]/[NeII] line flux ratios (or their upper limits) are significantly lower than implied by the MIR slope or EQW of the 6.2 micron PAH feature. Fitting the SEDs, we see evidence for hot (T > 300K) dust in five of the BGS ULIRGs, with the fraction of hot dust to total dust luminosity ranging from ~1-23%, before correcting for extinction. When integrated over the IRAC-8, IRS blue peakup, and MIPS-24 filter bandpasses, the IRS spectra imply very blue colors for some ULIRGs at z ~ 1.3. This is most extreme for sources with significant amounts of warm dust and deep silicate absorption.

astro-ph↗

arXiv2006

The Mid-IR Properties of Starburst Galaxies from Spitzer-IRS Spectroscopy

We present 5-38um mid-infrared spectra at a spectral resolution of R~65-130 of a large sample of 22 starburst nuclei taken with the Infrared Spectrograph IRS on board the Spitzer Space Telescope. The spectra show a vast range in starburst SEDs. The silicate absorption ranges from essentially no absorption to heavily obscured systems with an optical depth of tau(9.8um)~5. The spectral slopes can be used to discriminate between starburst and AGN powered sources. The monochromatic continuum fluxes at 15um and 30um enable a remarkably accurate estimate of the total infrared luminosity of the starburst. We find that the PAH equivalent width is independent of the total starburst luminosity L_IR as both continuum and PAH feature scale proportionally. However, the luminosity of the 6.2um feature scales with L_IR and can be used to approximate the total infrared luminosity of the starburst. Although our starburst sample covers about a factor of ten difference in the [NeIII]/[NeII] ratio, we found no systematic correlation between the radiation field hardness and the PAH equivalent width or the 7.7um/11.3um PAH ratio. These results are based on spatially integrated diagnostics over an entire starburst region, and local variations may be ``averaged out''. It is presumably due to this effect that unresolved starburst nuclei with significantly different global properties appear spectrally as rather similar members of one class of objects.

astro-ph↗

arXiv2006

Mid-Infrared Selection of Brown Dwarfs and High-Redshift Quasars

We discuss color selection of rare objects in a wide-field, multiband survey spanning from the optical to the mid-infrared. Simple color criteria simultaneously identify and distinguish two of the most sought after astrophysical sources: the coolest brown dwarfs and the most distant quasars. We present spectroscopically-confirmed examples of each class identified in the IRAC Shallow Survey of the Bootes field of the NOAO Deep Wide-Field Survey. ISS J142950.9+333012 is a T4.5 brown dwarf at a distance of approximately 42 pc, and ISS J142738.5+331242 is a radio-loud quasar at redshift z=6.12. Our selection criteria identify a total of four candidates over 8 square degrees of the Bootes field. The other two candidates are both confirmed 5.5<z<6 quasars, previously reported by Cool et al. (2006). We discuss the implications of these discoveries and conclude that there are excellent prospects for extending such searches to cooler brown dwarfs and higher redshift quasars.

astro-ph↗

arXiv2006

Photometric Redshifts in the IRAC Shallow Survey

Accurate photometric redshifts are calculated for nearly 200,000 galaxies to a 4.5 micron flux limit of ~13 uJy in the 8.5 deg^2 Spitzer/IRAC Shallow survey. Using a hybrid photometric redshift algorithm incorporating both neural-net and template-fitting techniques, calibrated with over 15,000 spectroscopic redshifts, a redshift accuracy of σ= 0.06(1+z) is achieved for 95% of galaxies at 0 1) galaxy clusters. We present one such spectroscopically confirmed cluster at =1.24, ISCS J1434.5+3427. Finally, we present a measurement of the 4.5 micron-selected galaxy redshift distribution.

astro-ph↗

arXiv2006

Spitzer IRS Spectra of Optically Faint Infrared Sources with Weak Spectral Features

Spectra have been obtained with the low-resolution modules of the Infrared Spectrograph (IRS) on the Spitzer Space Telescope (Spitzer) for 58 sources having f$_ν$(24 micron) > 0.75 mJy. Sources were chosen from a survey of 8.2 deg$^{2}$ within the NOAO Deep Wide-Field Survey region in Bootes (NDWFS) using the Multiband Imaging Photometer (MIPS) on the Spitzer Space Telescope. Most sources are optically very faint (I > 24mag). Redshifts have previously been determined for 34 sources, based primarily on the presence of a deep 9.7 micron silicate absorption feature, with a median z of 2.2. Spectra are presented for the remaining 24 sources for which we were previously unable to determine a confident redshift because the IRS spectra show no strong features. Optical photometry from the NDWFS and infrared photometry with MIPS and the Infrared Array Camera on the Spitzer Space Telescope (IRAC) are given, with K photometry from the Keck I telescope for some objects. The sources without strong spectral features have overall spectral energy distributions (SEDs) and distributions among optical and infrared fluxes which are similar to those for the sources with strong absorption features. Nine of the 24 sources are found to have feasible redshift determinations based on fits of a weak silicate absorption feature. Results confirm that the "1 mJy" population of 24 micron Spitzer sources which are optically faint is dominated by dusty sources with spectroscopic indicators of an obscured AGN rather than a starburst. There remain 14 of the 58 sources observed in Bootes for which no redshift could be estimated, and 5 of these sources are invisible at all optical wavelengths.

astro-ph↗

arXiv2006

A Spitzer Infrared Spectrograph Survey of Warm Molecular Hydrogen in Ultra-luminous Infrared Galaxies

We have conducted a survey of Ultra-luminous Infrared Galaxies (ULIRGs) with the Infrared Spectrograph on the Spitzer Space Telescope, obtaining spectra from 5.0-38.5um for 77 sources with 0.02 2Jy. The average warm molecular gas mass is ~2x10^8solar-masses. High extinction, inferred from the 9.7um silicate absorption depth, is not observed along the line of site to the molecular gas. The derived H2 mass does not depend on F(25um)/F(60um), which has been used to infer either starburst or AGN dominance. Similarly, the molecular mass does not scale with the 25 or 60um luminosities. In general, the H2 emission is consistent with an origin in photo-dissociation regions associated with star formation. We detect the S(0) 28.22um emission line in a few ULIRGs. Including this line in the model fits tends to lower the temperature by ~50-100K, resulting in a significant increase in the gas mass. The presence of a cooler component cannot be ruled out in the remainder of our sample, for which we do not detect the S(0) line. The measured S(7) 5.51um line fluxes in six ULIRGs implies ~3x10^6 solar-masses of hot (~1400K) H2. The warm gas mass is typically less than 1% of the cold gas mass derived from CO observations.

astro-ph↗