Abstract

We present an extensive CO emission-line survey of the Planck’s dusty Gravitationally Enhanced subMillimetre Sources, a small set of 11 strongly lensed dusty star-forming galaxies at z = 2–4 discovered with Planck and Herschel satellites, using EMIR on the IRAM 30-m telescope. We detected a total of 45 CO rotational lines from Jup = 3 to Jup = 11, and up to eight transitions per source, allowing a detailed analysis of the gas excitation and interstellar medium conditions within these extremely bright (μLFIR = 0.5 − 3.0 × 1014L⊙), vigorous starbursts. The peak of the CO spectral-line energy distributions (SLEDs) fall between Jup = 4 and Jup = 7 for nine out of 11 sources, in the same range as other lensed and unlensed submillimeter galaxies (SMGs) and the inner regions of local starbursts. We applied radiative transfer models using the large velocity gradient approach to infer the spatially-averaged molecular gas densities, nH2 ≃ 102.6 − 104.1 cm−3, and kinetic temperatures, Tk ≃ 30–1000 K. In five sources, we find evidence of two distinct gas phases with different properties and model their CO SLED with two excitation components. The warm (70–320 K) and dense gas reservoirs in these galaxies are highly excited, while the cooler (15–60 K) and more extended low-excitation components cover a range of gas densities. In two sources, the latter is associated with diffuse Milky Way-like gas phases of density nH2 ≃ 102.4 − 102.8 cm−3, which provides evidence that a significant fraction of the total gas masses of dusty starburst galaxies can be embedded in cool, low-density reservoirs. The delensed masses of the warm star-forming molecular gas range from 0.6to12 × 1010 M⊙. Finally, we show that the CO line luminosity ratios are consistent with those predicted by models of photon-dominated regions (PDRs) and disfavor scenarios of gas clouds irradiated by intense X-ray fields from active galactic nuclei. By combining CO, [C I] and [C II] line diagnostics, we obtain average PDR gas densities significantly higher than in normal star-forming galaxies at low-redshift, as well as far-ultraviolet radiation fields 102–104 times more intense than in the Milky Way. These spatially-averaged conditions are consistent with those in high-redshift SMGs and in a range of low-redshift environments, from the central regions of ultra-luminous infrared galaxies and bluer starbursts to Galactic giant molecular clouds.

Highlights

  • Massive, dusty, star-forming galaxies at high-redshift exhibit enhanced star-formation rates compared to those measured in low-redshift nuclear ultra-luminous infrared galaxies (ULIRGs), and account for significant fractions of the cosmic energy budget from star formation at z ∼ 2–4 (e.g., Hauser & Dwek 2001; Magnelli et al 2013; Dunlop et al 2017)

  • We found that the discrepancy vanishes when using the carbon monoxide (CO)(1–0) luminosities extrapolated from our large velocity gradient (LVG) models, except for PLCK_G165.7+67.0

  • Our density estimates from the LVG and photon-dominated regions (PDRs) models presented in Tables 3 and 4 differ by about one order of magnitude, and average PDR densities, nPDR = nH + 2 × nH2, are higher than nH2 measured on the integrated CO spectral line energy distributions (SLEDs)

Read more

Summary

Introduction

Dusty, star-forming galaxies at high-redshift exhibit enhanced star-formation rates (between a few hundred and 1000 M yr−1, for example, Casey et al 2014) compared to those measured in low-redshift nuclear ultra-luminous infrared galaxies (ULIRGs), and account for significant fractions of the cosmic energy budget from star formation at z ∼ 2–4 (e.g., Hauser & Dwek 2001; Magnelli et al 2013; Dunlop et al 2017) Their intense star formation is driven by deep gravitational potentials and higher gas masses (e.g., Tacconi et al 2008; Ivison et al 2011), gas fractions (e.g., Daddi et al 2010; Tacconi et al 2010), and gas mass surface densities (e.g., Riechers et al 2013; Cañameras et al 2017a) than in the local Universe. Observing the CO spectral line energy distributions (SLEDs) allows us to characterize the population of multiple rotational levels and to infer the underlying gas density and kinetic temperature using radiative transfer models (e.g., Weiß et al 2005; Papadopoulos 2010)

Methods
Findings
Discussion
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call