The energetic environment and the dense interstellar medium in ULIRGs

We fit the near-infrared to radio spectral energy distributions of a sample of 30 luminous and ultra-luminous infrared galaxies with models that include both starburst and AGN components. The aim of the work was to determine important physical parameters for this kind of objects such as the optical depth towards the luminosity source, the star formation rate, the star formation efficiency and the AGN fraction. We found that although about half of our sample have best-fit models that include an AGN component, only 30 % have an AGN which accounts for more than 10 % of the infrared luminosity whereas all have an energetically dominant starburst. Our models also determine the mass of dense molecular gas. Assuming that this mass is that traced by the HCN molecule, we reproduce the observed linear relation between HCN luminosity and infrared luminosity found by Gao and Solomon (2004). However, our derived conversion factor between HCN luminosity and the mass of dense molecular gas is a factor of 2 smaller than that assumed by these authors. Finally, we find that the star formation efficiency falls as the starburst ages.

💡 Research Summary

The paper presents a comprehensive study of the energetic drivers and dense interstellar medium (ISM) properties in a sample of 30 luminous (LIRG) and ultra‑luminous infrared galaxies (ULIRG). The authors assembled multi‑wavelength spectral energy distributions (SEDs) spanning from the near‑infrared (NIR) to the radio regime, drawing on data from facilities such as 2MASS, IRAS, Spitzer, Herschel, and the VLA. To interpret these SEDs they employed a hybrid model that simultaneously includes a starburst (SB) component and an active galactic nucleus (AGN) component. The SB component accounts for the emission of young stellar populations, dust absorption, and re‑radiation, while the AGN component models the dusty torus surrounding a supermassive black hole and its non‑thermal contribution. Free parameters include the visual optical depth (τ_V), star formation rate (SFR), starburst age (age_SB), star formation efficiency (SFE), and the fractional AGN contribution to the total infrared luminosity (f_AGN). Parameter space was explored using a Markov Chain Monte Carlo (MCMC) fitting routine, ensuring robust statistical uncertainties.

Key findings are as follows:

1. Dominance of Starbursts: Roughly half of the galaxies achieve their best‑fit SEDs with an AGN component, yet only about 30 % of the entire sample have an AGN that contributes more than 10 % of the total infrared luminosity (L_IR). Consequently, the bulk of the infrared output in ULIRGs is powered by intense star formation rather than by accretion onto a central black hole. This result challenges the notion that many ULIRGs are AGN‑dominated and reinforces the view that they are primarily starburst systems.

2. Dense Molecular Gas Mass and HCN Tracing: By extracting the mass of the dense molecular gas (M_dense) from the SED fits, the authors assume that this mass is traced by the HCN (1‑0) line. They successfully reproduce the linear relationship between HCN luminosity (L_HCN) and infrared luminosity reported by Gao & Solomon (2004). However, the conversion factor α_HCN = M_dense / L_HCN derived here is about a factor of two lower than the canonical value used by Gao & Solomon. This suggests that HCN may be tracing gas that is either denser, warmer, or more efficiently excited than previously thought, implying that existing estimates of dense gas mass in luminous infrared galaxies could be systematically overestimated.

3. Evolution of Star Formation Efficiency: The analysis reveals a clear trend of decreasing SFE with increasing starburst age. Young starbursts (age < 10 Myr) exhibit SFE values around 0.5 Gyr⁻¹, whereas more evolved systems (age > 30 Myr) show SFE dropping to ≲0.2 Gyr⁻¹. This decline is interpreted as a consequence of gas consumption and feedback processes (supernova explosions, stellar winds, radiation pressure) that progressively reduce the efficiency with which the remaining dense gas can be converted into new stars.

4. Optical Depth and AGN Visibility: A secondary correlation emerges between the visual optical depth and the observable AGN contribution. Galaxies with τ_V > 50 tend to have suppressed AGN fractions, indicating that thick dust columns effectively obscure the AGN’s infrared signature, making it difficult to detect even when an active nucleus is present.

Methodologically, the strength of the work lies in its use of broadband SED fitting, which mitigates biases that can arise when only a limited wavelength range is considered. By jointly fitting the SB and AGN components, the authors can disentangle their relative contributions with higher confidence than single‑component models.

The paper concludes by emphasizing three main implications for galaxy evolution studies: (i) ULIRGs are overwhelmingly starburst‑powered; (ii) the HCN‑to‑dense‑gas conversion factor needs revision, affecting mass estimates of the star‑forming reservoir; and (iii) star formation efficiency is not static but declines as the starburst ages, reflecting the evolving physical conditions within the ISM. The authors suggest that future high‑resolution observations with ALMA and JWST, combined with sophisticated radiative‑transfer modeling, will be essential to map the three‑dimensional structure of the dusty torus and starburst regions, thereby refining our understanding of the interplay between star formation and AGN activity in the most luminous infrared galaxies.