Investigating ultraviolet and infrared radiation through the turbulent life of molecular clouds

Context. Molecular clouds (MCs) are the places where stars are formed and their feedback starts to take place, regulating the evolution of galaxies. Therefore, MCs represent the critical scale at which to study how ultraviolet (UV) photons emitted by young stars are reprocessed in the far-infrared (FIR) by interaction with dust grains, thereby determining the multiwavelength continuum emission of galaxies. Aims. Our goal is to analyze the UV and IR emission of a MC at different stages of its evolution and relate its absorption and emission properties with its morphology and star formation rate. Such a study is fundamental to determining how the properties of MCs shape the emission from entire galaxies. Methods. We considered a radiation-hydrodynamic simulation of a MC with self-consistent chemistry treatment. The MC has a mass of M MC = 10 5 M ⊙ , is resolved down to a scale of 0.06 pc, and evolves for ≃2.4 Myr after the onset of star formation. We post-processed the simulation via Monte Carlo radiative transfer calculations to compute the detailed UV-to-FIR emission of the MC. Such results were compared with data from physically motivated analytical models, other simulations, and observations. Results. We find that the simulated MC is globally UV-optically thick, but optically thin channels allow for photon escape (0.1–10%), a feature that is not well captured in analytical models. The dust temperature spans a wide range ( T dust ∼ 20–300 K) depending on the dust-to-stellar geometry, which is reproduced reasonably well by analytical models. However, the complexity of the dust temperature distribution is not captured in the analytical models, as is evidenced by the 10 K (20 K) difference in the mass (luminosity) average temperature. Indeed, the total IR luminosity is the same in all the models, but the IR emission peaks at shorter wavelengths in the analytical ones. Compared to a sample of Galactic clouds and other simulations, our spectral energy distribution (SED) is consistent with mid-IR data, but peaks at shorter wavelengths in the IR. This is due to a lack of cold dust, as a consequence of the high gas – and thus dust – consumption in our simulated MC. The attenuation properties of our MC change significantly with time, evolving from a Milky-Way-like relation to a flatter, featureless one. On the IRX-β plane, the MC position strongly depends on the observing direction and on its evolutionary stage. When the MC starts to disperse, the cloud settles at log(IRX) ∼ 1 and β ∼ −0.5, slightly below most of the local empirical relations. Conclusions. This work represents an important test for MC simulations and a first step toward the implementation of a physically informed, sub-grid model in large-scale numerical simulations to describe the emission from unresolved MC scales and its impact on the global galaxy SED.