SIMBAD references

Context. Previous studies have shown the existence of an excess of emission at submillimeter (submm) and millimeter (mm) wavelengths in the spectral energy distribution (SED) of many low-metallicity galaxies. The so-called "submm excess", whose origin remains unknown, challenges our understanding of the dust properties in low-metallicity environments.
Aims. The goal of the present study is to model separately the emission from the star forming (SF) component and the emission from the diffuse interstellar medium (ISM) in the nearby spiral galaxy M 33 in order to determine whether both components can be well fitted using radiation transfer models or whether there is an excess of submm emission associated with one or both of them.
Methods. We decomposed the observed SED of M 33 into its SF and diffuse components. Mid-infrared (MIR) and far-infrared (FIR) fluxes were extracted from Spitzer and Herschel data. At submm and mm wavelengths, we used ground-based observations from APEX to measure the emission from the SF component and data from the Planck space telescope to estimate the diffuse emission. Both components were separately fitted using radiation transfer models based on standard dust properties (i.e., emissivity index β=2) and a realistic geometry. The large number of previous studies helped us to estimate the thermal radio emission and to constrain an important part of the input parameters of the models. Both modeled SEDs were combined to build the global SED of M 33. In addition, the radiation field necessary to power the dust emission in our modeling was compared with observations from GALEX, Sloan, and Spitzer.
Results. Our modeling is able to reproduce the observations at MIR and FIR wavelengths, but we found a strong excess of emission at submm and mm wavelengths where the model expectations severely underestimate the LABOCA and Planck fluxes. We also found that the ultraviolet (UV) radiation escaping the galaxy is 70% higher than the model predictions. From the total mass of dust derived from our modeling and the mass of atomic and molecular gas measured with the VLA and the IRAM 30m telescope, we determined a gas-to-dust mass ratio G_dust∼100, significantly lower than the value expected from the subsolar metallicity of M 33.
Conclusions. We discussed different hypotheses to explain the discrepancies found in our study (i.e., excess of emission at submm and mm wavelengths, deficit of UV attenuation, and abnormally low value of G_dust), concluding that different dust properties in M 33 is the most plausible explanation.