arXiv subjects

L. Colina

Publications and source records attributed to L. Colina.

At least 109 records · Page 6Linked to original sources

arXiv2014

Molecular line emission in NGC1068 imaged with ALMA. I An AGN-driven outflow in the dense molecular gas

We investigate the fueling and the feedback of star formation and nuclear activity in NGC1068, a nearby (D=14Mpc) Seyfert 2 barred galaxy, by analyzing the distribution and kinematics of the molecular gas in the disk. We have used ALMA to map the emission of a set of dense molecular gas tracers (CO(3-2), CO(6-5), HCN(4-3), HCO+(4-3) and CS(7-6)) and their underlying continuum emission in the central r ~ 2kpc of NGC1068 with spatial resolutions ~ 0.3"-0.5" (~ 20-35pc). Molecular line and dust continuum emissions are detected from a r ~ 200pc off-centered circumnuclear disk (CND), from the 2.6kpc-diameter bar region, and from the r ~ 1.3kpc starburst (SB) ring. Most of the emission in HCO+, HCN and CS stems from the CND. Molecular line ratios show dramatic order-of-magnitude changes inside the CND that are correlated with the UV/X-ray illumination by the AGN, betraying ongoing feedback. The gas kinematics from r ~ 50pc out to r ~ 400pc reveal a massive (M_mol ~ 2.7 (+0.9, -1.2) x 10^7 Msun) outflow in all molecular tracers. The tight correlation between the ionized gas outflow, the radio jet and the occurrence of outward motions in the disk suggests that the outflow is AGN-driven. The outflow rate estimated in the CND, dM/dt ~ 63 (+21, -37) Msun yr^-1, is an order of magnitude higher than the star formation rate at these radii, confirming that the outflow is AGN-driven. The power of the AGN is able to account for the estimated momentum and kinetic luminosity of the outflow. The CND mass load rate of the CND outflow implies a very short gas depletion time scale of <=1 Myr.

astro-ph.GA↗

arXiv2013

Uncovering the deeply embedded AGN activity in the nuclear regions of the interacting galaxy Arp299

We present mid-infrared (MIR) 8-13micron spectroscopy of the nuclear regions of the interacting galaxy Arp299 (IC694+NGC3690) obtained with CanariCam (CC) on the 10.4m Gran Telescopio Canarias (GTC). The high angular resolution (~0.3-0.6arcsec) of the data allows us to probe nuclear physical scales between 60 and 120pc, which is a factor of 10 improvement over previous MIR spectroscopic observations of this system. The GTC/CC spectroscopy displays evidence of deeply embedded Active Galactic Nucleus (AGN) activity in both nuclei. The GTC/CC nuclear spectrum of NGC3690/Arp299-B1 can be explained as emission from AGN-heated dust in a clumpy torus with both a high covering factor and high extinction along the line of sight. The estimated bolometric luminosity of the AGN in NGC3690 is 3.2(+/-0.6)x10^44 erg/s. The nuclear GTC/CC spectrum of IC694/Arp299-A shows 11.3micron polycyclic aromatic hydrocarbon (PAH) emission stemming from a deeply embedded (A_V~24mag) region of less than 120pc in size. There is also a continuum-emitting dust component. If associated with th putative AGN in IC694, we estimate that it would be approximately 5 times less luminous than the AGN in NGC3690. The presence of dual AGN activity makes Arp299 a good example to study such phenomenon in the early coalescence phase of interacting galaxies.

astro-ph.CO↗

arXiv2013

Molecular gas in type 2 quasars at z~0.2-0.3

We present results of CO(1-0) spectroscopic observations of 10 SDSS type 2 quasars (QSO2) at z~0.2-0.3 observed with the 30m IRAM radiotelescope and the Australia Telescope Compact Array. We report 5 new confirmed CO(1-0) detections and 1 tentative detection. They have L'(CO)~several x 1e9 K km s-1 pc^2, while upper limits for the non detections are L'(CO)<~ several x 1e9 K km s-1 pc^2. This study increases the total number of QSO2 with CO measurements at z<~1 to 20, with a 50% detection rate.

astro-ph.CO↗

arXiv2013

VLT-SINFONI integral field spectroscopy of low-z luminous and ultraluminous infrared galaxies II. 2D extinction structure and distance effects

We present a 2D study of the internal extinction on (sub)kiloparsec scales of a sample of local (z<0.1) (U)LIRGs, based on near-infrared Pa_a, Br_d, and Br_g line ratios, obtained with VLT-SINFONI integral-field spectroscopy. The 2D extinction (Av) distributions of the objects, map regions of ~3x3 kpc (LIRGs) and ~12x12 kpc (ULIRGs), with average angular resolutions (FWHM) of ~0.2 kpc and ~0.9 kpc, respectively. The individual Av galaxy distributions indicate a very clumpy dust structure already on sub-kiloparsec scales, with values (per spaxel) ranging from Av ~ 1 to 20 mag in LIRGs, and from Av ~ 2 to 15 mag in ULIRGs. As a class, the median values of the distributions are Av=5.3 mag and Av=6.5 mag for the LIRG and ULIRG subsamples, respectively. We evaluated the effects of the galaxy distance in the measurements of the extinction as a function of the linear scale (in kpc) of the spaxel (i.e. due to the limited angular resolution of the observations). If the distribution of the gas/dust and star-forming regions in local LIRGs (63 Mpc, 40 pc/spaxel on average) is the same for galaxies at greater distances, the observed median Av values based on emission line ratios would be a factor ~0.8 lower at the average distance of our ULIRG sample (328 Mpc, 0.2 kpc/spaxel), and a factor ~0.67 for galaxies located at distances of more than 800 Mpc (0.4 kpc/spaxel). This distance effect would have implications for deriving the intrinsic extinction in high-z star-forming galaxies and for subsequent properties such as star formation rate, star formation surface density, and KS- law, based on H_alpha line fluxes. If local LIRGs are analogues of the main-sequence star-forming galaxies at cosmological distances, the extinction values (Av) derived from the observed emission lines in these high-z sources would need to be increased by a factor 1.4 on average.

astro-ph.CO↗

arXiv2012

Gas fraction and star formation efficiency at z < 1.0

After new observations of 39 galaxies at z = 0.6-1.0 obtained at the IRAM 30m telescope, we present our full CO line survey covering the redshift range 0.2 < z < 1. Our aim is to determine the driving factors accounting for the steep decline in the star formation rate during this epoch. We study both the gas fraction, defined as Mgas/(Mgas+Mstar), and the star formation efficiency (SFE) defined by the ratio between far-infrared luminosity and molecular gas mass (LFIR/M(H2), i.e. a measure for the inverse of the gas depletion time. The sources are selected to be ultra-luminous infrared galaxies (ULIRGs), with LFIR greater than 10^12 Lo and experiencing starbursts. When we adopt a standard ULIRG CO-to-H2 conversion factor, their molecular gas depletion time is less than 100 Myr. Our full survey has now filled the gap of CO observations in the 0.2<z<1 range covering almost half of cosmic history. The detection rate in the 0.6 < z < 1 interval is 38% (15 galaxies out of 39), compared to 60% for the 0.2<z<0.6 interval. The average CO luminosity is L'CO = 1.8 10^10 K km/s pc^2, corresponding to an average H2 mass of 1.45 10^10 Mo. From observation of 7 galaxies in both CO(2-1) and CO(4-3), a high gas excitation has been derived; together with the dust mass estimation, this supports the choice of our low ULIRG conversion factor between CO luminosity and H2, for our sample sources. We find that both the gas fraction and the SFE significantly increase with redshift, by factors of 3 +-1 from z=0 to 1, and therefore both quantities play an important role and complement each other in cosmic star formation evolution.

astro-ph.CO↗

arXiv2012

Star Formation Efficiency at Intermediate Redshift

Star formation is evolving very fast in the second half of the Universe, and it is yet unclear whether this is due to evolving gas content, or evolving star formation efficiency (SFE). We have carried out a survey of ultra-luminous galaxies (ULIRG) between z=0.2 and 1, to check the gas fraction in this domain of redshift which is still poorly known. Our survey with the IRAM-30m detected 33 galaxies out of 69, and we derive a significant evolution of both the gas fraction and SFE of ULIRGs over the whole period, and in particular a turning point around z=0.35. The result is sensitive to the CO-to-H2, conversion factor adopted, and both gas fraction and SFE have comparable evolution, when we adopt the low starburst conversion factor of α=0.8 Mo/(K km/s pc^2). Adopting a higher αwill increase the role of the gas fraction. Using α=0.8, the SFE and the gas fraction for z=0.2-1.0 ULIRGs are found to be significantly higher, by a factor 3, than for local ULIRGs, and are comparable to high redshift ones. We compare this evolution to the expected cosmic H2 abundance and the cosmic star formation history.

astro-ph.CO↗

arXiv2012

VLT-SINFONI integral field spectroscopy of low-z luminous and ultraluminous infrared galaxies I. Atlas of the 2D gas structure

We present an atlas of a sample of local (U)LIRGs covering the luminosity range log(L_IR/L_sun)=11.1-12.4. The atlas is based on near-infrared H and K-band VLT-SINFONI IFS, and presents the ionised, partially ionised, and warm molecular gas 2D flux distributions and kinematics over a FoV of 3x3 kpc (LIRGs) and 12x12kpc (ULIRGs) and with average linear resolutions of 0.2kpc and 0.9kpc, respectively. The different phases of the gas show a wide morphological variety with the nucleus as the brightest Br_g source for 33% of the LIRGs and 71% of the ULIRGs, whereas all the (U)LIRGs have their maximum H_2 emission in their nuclear regions. In LIRGs, the ionised gas distribution is dominated by the emission from the star-forming rings or giant HII regions in the spiral arms. The Br_g and [FeII] line at 1.644 micron trace the same structures, although the emission peaks at different locations in some of the objects, and the [FeII] seems to be more extended and diffuse. The ULIRG subsample contains mainly pre-coalescence interacting systems. Although the peaks of the molecular gas emission and the continuum coincide in 71% of the ULIRGs, regions with intense Pa_a (Br_g) emission tracing luminous star-forming regions located at distances of 2-4kpc away from the nucleus are also detected, usually associated with secondary nuclei or tidal tails. The gas kinematics in LIRGs are primarily due to rotational motions around the centre of the galaxy, although local deviations associated with radial flows and/or regions of higher velocity dispersions are present. The ionised and molecular gas share the same kinematics to first order, showing slight differences in the velocity amplitudes in some cases, whereas the average velocity dispersions are compatible within uncertainties. As expected, the kinematics of the ULIRG subsample is more complex, owing to the interacting nature of the objects of the sample.

astro-ph.CO↗

arXiv2012

The Nuclear Infrared Emission of Low-Luminosity Active Galactic Nuclei

We present high-resolution mid-infrared (MIR) imaging, nuclear spectral energy distributions (SEDs) and archival Spitzer spectra for 22 low-luminosity active galactic nuclei (LLAGN; Lbol \lesssim 10^42 erg/sec). Infrared (IR) observations may advance our understanding of the accretion flows in LLAGN, the fate of the obscuring torus at low accretion rates, and, perhaps, the star formation histories of these objects. However, while comprehensively studied in higher-luminosity Seyferts and quasars, the nuclear IR properties of LLAGN have not yet been well-determined. We separate the present LLAGN sample into three categories depending on their Eddington ratio and radio emission, finding different IR characteristics for each class. (I) At the low-luminosity, low-Eddington ratio (log Lbol/LEdd < -4.6) end of the sample, we identify "host-dominated" galaxies with strong polycyclic aromatic hydrocarbon bands that may indicate active (circum-)nuclear star formation. (II) Some very radio-loud objects are also present at these low Eddington ratios. The IR emission in these nuclei is dominated by synchrotron radiation, and some are likely to be unobscured type 2 AGN that genuinely lack a broad line region. (III) At higher Eddington ratios, strong, compact nuclear sources are visible in the MIR images. The nuclear SEDs of these galaxies are diverse; some resemble typical Seyfert nuclei, while others lack a well-defined MIR "dust bump". Strong silicate emission is present in many of these objects. We speculate that this, together with high ratios of silicate strength to hydrogen column density, could suggest optically thin dust and low dust-to-gas ratios, in accordance with model predictions that LLAGN do not host a Seyfert-like obscuring torus.

astro-ph.CO↗

arXiv2012

Spatially resolved kinematics of the central regions of M83: hidden mass signatures and the role of supernovae

The barred grand-design spiral M83 (NGC 5236) is one of the most studied galaxies given its proximity, orientation, and particular complexity. Nonetheless, many aspects of the central regions remain controversial conveying our limited understanding of the inner gas and stellar kinematics, and ultimately of the nucleus evolution. In this work, we present AO VLT-SINFONI data of its central ~235x140 pc with an unprecedented spatial resolution of ~0.2 arcsec, corresponding to ~4 pc. We have focused our study on the distribution and kinematics of the stars and the ionised and molecular gas by studying in detail the Pa_alpha and Br_gamma emission, the H_2 1-0S(1) line at 2.122 micron and the [FeII] line at 1.644 micron, together with the CO absorption bands at 2.293 micron and 2.323 micron. Our results reveal a complex situation where the gas and stellar kinematics are totally unrelated. Supernova explosions play an important role in shaping the gas kinematics, dominated by shocks and inflows at scales of tens of parsecs that make them unsuitable to derive general dynamical properties. We propose that the location of the nucleus of M83 is unlikely to be related to the off-centre 'optical nucleus'. The study of the stellar kinematics reveals that the optical nucleus is a gravitationally bound massive star cluster with M_dyn = (1.1 \pm 0.4)x10^7 M_sun, formed by a past starburst. The kinematic and photometric analysis of the cluster yield that the stellar content of the cluster is well described by an intermediate age population of log T(yr) = 8.0\pm0.4, with a mass of M \simeq (7.8\pm2.4)x10^6 M_sun.

astro-ph.CO↗

arXiv2012

Integral Field Spectroscopy based Hα sizes of local Luminous and Ultraluminous Infrared Galaxies. A Direct Comparison with high-z Massive Star Forming Galaxies

Aims. We study the analogy between local U/LIRGs and high-z massive SFGs by comparing basic Hα structural characteristics, such as size, and luminosity (and SFR) surface density, in an homogeneous way (i.e. same tracer and size definition, similar physical scales). Methods. We use Integral Field Spectroscopy based Hα emission maps for a representative sample of 54 local U/LIRGs (66 galaxies). From this initial sample we select 26 objects with Hα luminosities (L(Hα)) similar to those of massive (i.e. M\ast \sim 10^10 M\odot or larger) SFGs at z \sim 2, and observed on similar physical scales. Results. The sizes of the Hα emitting region in the sample of local U/LIRGs span a large range, with r1/2(Hα) from 0.2 to 7 kpc. However, about 2/3 of local U/LIRGs with Lir > 10^11.4 L\odot have compact Hα emission (i.e. r1/2 < 2 kpc). The comparison sample of local U/LIRGs also shows a higher fraction (59%) of objects with compact Hα emission than the high-z sample (25%). This gives further support to the idea that for this luminosity range the size of the star forming region is a distinctive factor between local and distant galaxies of similar SF rates. However, when using Hα as a tracer for both local and high-z samples, the differences are smaller than the ones recently reported using a variety of other tracers. Despite of the higher fraction of galaxies with compact Hα emission, a sizable group (\sim 1/3) of local U/LIRGs are large (i.e. r1/2 > 2 kpc). These are systems showing pre-coalescence merger activity and they are indistinguishable from the massive high-z SFGs galaxies in terms of their Hα sizes, and luminosity and SFR surface densities.

astro-ph.CO↗

arXiv2012

Studying the kinematic asymmetries of disks and post-coalescence mergers using a new `kinemetry' criterion

We have obtained VIMOS/VLT optical integral field spectroscopy (IFS) data for a sample of 4 LIRGs which have been selected at a similar distance ($\sim$ 70 Mpc) to avoid relative resolution effects. They have been classified in two groups (isolated disk and post-coalescence mergers) according to their morphology. The $kinemetry$ method (developed by Krajnovic and coworkers) is used to characterize the kinematic properties of these galaxies and to discuss new criteria for distinguishing their status. We present and discuss new kinematic maps (i.e., velocity field and velocity dispersion) for these four galaxies. The morphological and kinematic classifications of these systems are consistent, with disks having lower kinematic asymmetries than post-coalescence mergers. We then propose and discuss a new kinematic criterion to differentiate these two groups. This criterion distinguishes better these two categories and has the advantage of being less sensitive to angular resolution effects. According to the previous criteria,the present post-coalescence systems would have been classified as disks. This indicates that the separation of disks from mergers is subjective to the definition of `merger'. It also suggests that previous estimates of the merger/disk ratio could have been underestimated, but larger samples are necessary to establish a firmer conclusion.

astro-ph.CO↗

arXiv2012

Spatially Resolved [FeII] 1.64 μm Emission in NGC 5135. Clues for Understanding the Origin of the Hard X-rays in Luminous Infrared Galaxies

Spatially resolved near-IR and X-ray imaging of the central region of the Luminous Infrared Galaxy NGC 5135 is presented. The kinematical signatures of strong outflows are detected in the [FeII]1.64 μm emission line in a compact region at 0.9 kpc from the nucleus. The derived mechanical energy release is consistent with a supernova rate of 0.05-0.1 yr$^{-1}$. The apex of the outflowing gas spatially coincides with the strongest [FeII] emission peak and with the dominant component of the extranuclear hard X-ray emission. All these features provide evidence for a plausible direct physical link between supernova-driven outflows and the hard X-ray emitting gas in a LIRG. This result is consistent with model predictions of starbursts concentrated in small volumes and with high thermalization efficiencies. A single high-mass X-ray binary (HMXB) as the major source of the hard X-ray emission although not favoured, cannot be ruled out. Outside the AGN, the hard X-ray emission in NGC 5135 appears to be dominated by the hot ISM produced by supernova explosions in a compact star-forming region, and not by the emission due to HMXB. If this scenario is common to U/LIRGs, the hard X-rays would only trace the most compact (< 100 pc) regions with high supernova and star formation densities, therefore a lower limit to their integrated star formation. The SFR derived in NGC 5135 based on its hard X-ray luminosity is a factor of two and four lower than the values obtained from the 24 μm and soft X-ray luminosities, respectively.

astro-ph.CO↗

arXiv2011

Integrated spectra extraction based on signal-to-noise optimization using Integral Field Spectroscopy

We propose and explore the potential of a method to extract high signal-to-noise (S/N) integrated spectra related to physical and/or morphological regions on a 2-dimensional field using Integral Field Spectroscopy (IFS) observations by employing an optimization procedure based on either continuum (stellar) or line (nebular) emission features. The optimization method is applied to a set of IFS VLT-VIMOS observations of (U)LIRG galaxies, describing the advantages of the optimization by comparing the results with a fixed-aperture, single spectrum case, and by implementing some statistical tests. We demonstrate that the S/N of the IFS optimized integrated spectra is significantly enhanced when compared with the single aperture unprocessed case. We provide an iterative user-friendly and versatile IDL algorithm that allows the user to spatially integrate spectra following more standard procedures. This is made available to the community as part of the PINGSoft IFS software package.

astro-ph.IM↗

arXiv2011

Discovery of Two Supernovae in the Nuclear Regions of the Luminous Infrared Galaxy IC 883

We report the discovery of two consecutive supernovae (SNe), 2010cu and 2011hi, located at 0.37" (180 pc) and 0.79" (380 pc) projected distance respectively from the centre of the K-band nucleus of the luminous infrared galaxy IC 883. The SNe were discovered in an ongoing near-infrared K-band search for core-collapse SNe in such galaxies using the ALTAIR/NIRI adaptive optics system with laser guide star at the Gemini-North Telescope. These are thus the closest SNe yet discovered to a LIRG nucleus in optical or near-infrared wavelengths. The near-infrared light curves and colours of both SNe are consistent with core-collapse events. Both SNe seem to suffer from relatively low host galaxy extinction suggesting that regardless of their low projected galactocentric distances, they are not deeply buried in the nuclear regions of the host galaxy.

astro-ph.SR↗

arXiv2011

Star formation laws in Luminous Infrared Galaxies. New observational constraints on models

The observational study of star formation relations in galaxies is central to unraveling the physical processes at work on local and global scales. We wish to expand the sample of extreme starbursts, represented by local LIRGs and ULIRGs, with high quality observations in the 1-0 line of HCN. We study if a universal law can account for the star formation relations observed for the dense molecular gas in normal star forming galaxies and extreme starbursts. We have used the IRAM 30m telescope to observe a sample of 19 LIRGs in the 1-0 lines of CO, HCN and HCO+. The analysis of the new data proves that the efficiency of star formation in the dense molecular gas (SFE-dense) of extreme starbursts is a factor 3-4 higher compared to normal galaxies. We find a duality in Kennicutt-Schmidt (KS) laws that is reinforced if we account for the different conversion factor for HCN (alpha-HCN) in extreme starbursts and for the unobscured star formation rate in normal galaxies. This result extends to the higher molecular densities probed by HCN lines the more extreme bimodal behavior of star formation laws, derived from CO molecular lines by two recent surveys. We have confronted our observations with the predictions of theoretical models in which the efficiency of star formation is determined by the ratio of a constant star formation rate per free-fall time (SFR-ff) to the local free-fall time. We find that it is possible to fit the observed differences in the SFE-dense between normal galaxies and LIRGs/ULIRGs using a common constant SFR-ff and a set of physically acceptable HCN densities, but only if SFR-ff~0.005-0.01 and/or if alpha-HCN is a factor of a few lower than our favored values. Star formation recipes that explicitly depend on the galaxy global dynamical time scales do not significantly improve the fit to the new HCN data presented in this work.

astro-ph.CO↗

arXiv2011

Extranuclear Halpha-emitting complexes in low-z (U)LIRGs: Precursors of tidal dwarf galaxies?

(Abridged)This paper characterizes the physical and kinematic properties of external massive star-forming regions in a sample of (U)LIRGs. We use high angular resolution ACS images from the HST B and I bands, as well as Halpha-line emission maps obtained with IFS. We find 31 external Halpha-emitting (young star-forming) complexes in 11 (U)LIRGs. These complexes have in general similar sizes, luminosities, and metallicities to extragalactic giant HII regions and TDG candidates found in less luminous mergers and compact groups of galaxies. We assess the mass content and the likelihood of survival as TDGs of the 22 complexes with simple structures in the HST images based on their photometric, structural, and kinematic properties. The dynamical tracers used (radius-sigma and luminosity-sigma diagrams) indicate that most of the complexes might be self-gravitating entities. The resistance to forces from the parent galaxy is studied by considering the tidal mass of the candidate and its relative velocity with respect to the parent galaxy. After combining the results of previous studies of TDG searches in ULIRGs a total of 9 complexes satisfy most of the applied criteria and thus show a high-medium or high likelihood of survival, their total mass likely being compatible with that of dwarf galaxies. They are defined as TDG candidates. We propose that they probably formed more often during the early phases of the interaction. Combining all data for complexes with IFS data where a significant fraction of the system is covered, we infer a TDG production rate of 0.3 candidates with the highest probabilities of survival per system for the (U)LIRGs class. This rate, though, might decrease to 0.1 after the systems in (U)LIRGs have evolved for 10 Gyr, for long-lived TDGs, which would imply that no more than 5-10 % of the overall dwarf population could be of tidal origin.

astro-ph.CO↗

arXiv2011

Ionized outflows in SDSS type 2 quasars at z~0.3-0.6

We have analyzed the spatially integrated kinematic properties of the ionized gas within the inner r ~2). While the quiescent gas shows small or null velocity shifts relative to the systemic velocity, the highly perturbed gas trends to show larger shifts which, moreover, are blueshifts in general. Within a given object, the most perturbed gas trends to have the largest blueshift as well. All together support that the perturbed gas, which is responsible for the blue asymmetries of the line profiles, is outflowing. Although some bias affects the sample, we argue that ionized gas outflows are a common phenomenon in optically selected type 2 quasars at 0.3< z < 0.6.

astro-ph.CO↗

arXiv2011

Galaxy Evolution and Star Formation Efficiency at 0.2 < z < 0.6

We present the results of a CO line survey of 30 galaxies at moderate redshift (z \sim 0.2-0.6), with the IRAM 30m telescope, with the goal to follow galaxy evolution and in particular the star formation efficiency (SFE) as defined by the ratio between far-infrared luminosity and molecular gas mass (LFIR/M(H2)). The sources are selected to be ultra-luminous infrared galaxies (ULIRGs), with LFIR larger than 2.8 10^{12} Lsol, experiencing starbursts; adopting a low ULIRG CO-to-H2 conversion factor, their gas consumption time-scale is lower than 10^8 yr. To date only very few CO observations exist in this redshift range that spans nearly 25% of the universe's age. Considerable evolution of the star formation rate is already observed during this period. 18 galaxies out of our sample of 30 are detected (of which 16 are new detections), corresponding to a detection rate of 60%. The average CO luminosity for the 18 galaxies detected is L'CO = 2 10^{10} K km/s pc^2, corresponding to an average H2 mass of 1.6 10^{10} Msol. The FIR luminosity correlates well with the CO luminosity, in agreement with the correlation found for low and high redshift ULIRGs. Although the conversion factor between CO luminosity and H2 mass is uncertain, we find that the maximum amount of gas available for a single galaxy is quickly increasing as a function of redshift. Using the same conversion factor, the SFEs for z\sim 0.2-0.6 ULIRGs are found to be significantly higher, by a factor 3, than for local ULIRGs, and are comparable to high redshift ones. We compare this evolution to the expected cosmic H2 abundance and the cosmic star formation history.

astro-ph.CO↗