Aims. The goal of this paper is to analyse the behaviour of the gas-to-dust mass ratio (G/D) of local Universe galaxies over a wide metallicity range. We especially focus on the low-metallicity part of the G/D vs metallicity relation and investigate several explanations for the observed relation and scatter. Methods. We assembled a total of 126 galaxies, covering a 2 dex metallicity range and with 30% of the sample with 12 + log(O/H) ≤ 8.0. We homogeneously determined the dust masses with a semi-empirical dust model including submm constraints. The atomic and molecular gas masses have been compiled from the literature. We used two X CO scenarios to estimate the molecular gas mass: the Galactic conversion factor, X CO,MW , and a X CO that depends on the metallicity X CO,Z (∝Z −2 ). We modelled the observed trend of the G/D with metallicity using two simple power laws (slope of -1 and free) and a broken power law. Correlations with morphological type, stellar masses, star formation rates, and specific star formation rates are also discussed. We then compared the observed evolution of the G/D with predictions from several chemical evolution models and explored different physical explanations for the observed scatter in the G/D values. Results. We find that out of the five tested galactic parameters, metallicity is the main physical property of the galaxy driving the observed G/D. The G/D versus metallicity relation cannot be represented by a single power law with a slope of -1 over the whole metallicity range. The observed trend is steeper for metallicities lower than ∼8.0. A large scatter is observed in the G/D values for a given metallicity: in metallicity bins of ∼0.1 dex, the dispersion around the mean value is ∼0.37 dex. On average, the broken power law reproduces the observed G/D best compared to the two power laws (slope of -1 or free) and provides estimates of the G/D that are accurate to a factor of 1.6. The good agreement of observed values of the G/D and its scatter with respect to metallicity with the predicted values of the three tested chemical evolution models allows us to infer that the scatter in the relation is intrinsic to galactic properties, reflecting the different star formation histories, dust destruction efficiencies, dust grain size distributions, and chemical compositions across the sample. Conclusions. Our results show that the chemical evolution of low-metallicity galaxies, traced by their G/D, strongly depends on their local internal conditions and individual histories. The large scatter in the observed G/D at a given metallicity reflects the impact of various processes occurring during the evolution of a galaxy. Despite the numerous degeneracies affecting them, disentangling these various processes is now the next step.

Context. The far-infrared (FIR) lines are important tracers of the cooling and physical conditions of the interstellar medium (ISM) and are rapidly becoming workhorse diagnostics for galaxies throughout the universe. There are clear indications of a different behavior of these lines at low metallicity that needs to be explored. Aims. Our goal is to explain the main differences and trends observed in the FIR line emission of dwarf galaxies compared to more metal-rich galaxies, and how this translates in ISM properties. ] 57 µm fine-structure cooling lines in a sample of 48 low-metallicity star-forming galaxies of the guaranteed time key program Dwarf Galaxy Survey. We correlate PACS line ratios and line-to-L TIR ratios with L TIR , L TIR /L B , metallicity, and FIR color, and interpret the observed trends in terms of ISM conditions and phase filling factors with Cloudy radiative transfer models. Results. We find that the FIR lines together account for up to 3 percent of L TIR and that star-forming regions dominate the overall emission in dwarf galaxies. Compared to metal-rich galaxies, the ratios of Harboring compact phases of a low filling factor and a large volume filling factor of diffuse gas, the ISM of low-metallicity dwarf galaxies has a more porous structure than that of metal-rich galaxies. Methods

Aims. We analyze the applicability of far-infrared fine-structure lines [C] 158 µm, [O] 63 µm, and [O] 88 µm to reliably trace the star formation rate (SFR) in a sample of low-metallicity dwarf galaxies from the Herschel Dwarf Galaxy Survey and, furthermore, extend the analysis to a broad sample of galaxies of various types and metallicities in the literature. Methods. We study the trends and scatter in the relation between the SFR (as traced by GALEX FUV and MIPS 24 µm) and farinfrared line emission, on spatially resolved and global galaxy scales, in dwarf galaxies. We assemble far-infrared line measurements from the literature and infer whether the far-infrared lines can probe the SFR (as traced by the total infrared luminosity) in a variety of galaxy populations. Results. In metal-poor dwarfs, the [O] 63 and [O] 88 lines show the strongest correlation with the SFR with an uncertainty on the SFR estimates better than a factor of 2, while the link between [C] emission and the SFR is more dispersed (uncertainty factor of 2.6). The increased scatter in the SFR-L [CII] relation toward low metal abundances, warm dust temperatures, and large filling factors of diffuse, highly ionized gas suggests that other cooling lines start to dominate depending on the density and ionization state of the gas. For the literature sample, we evaluate the correlations for a number of different galaxy populations. The [C] and [O] 63 lines are considered to be reliable SFR tracers in starburst galaxies, recovering the star formation activity within an uncertainty of factor 2. For sources with composite and active galactic nucleus (AGN) classifications, all three FIR lines can recover the SFR with an uncertainty factor of 2.3. The SFR calibrations for ultra-luminous infrared galaxies (ULIRGs) are similar to starbursts/AGNs in terms of scatter but offset from the starburst/AGN SFR relations because of line deficits relative to their total infrared luminosity. While the number of detections of the FIR fine-structure lines is still very limited at high redshift for [O]

Context. We present new photometric data from our Herschel guaranteed time key programme, the Dwarf Galaxy Survey (DGS), dedicated to the observation of the gas and dust in low-metallicity environments. A total of 48 dwarf galaxies were observed with the PACS and SPIRE instruments onboard the Herschel Space Observatory at 70, 100, 160, 250, 350, and 500 µm. Aims. The goal of this paper is to provide reliable far-infrared (FIR) photometry for the DGS sample and to analyse the FIR/submillimetre (submm) behaviour of the DGS galaxies. We focus on a systematic comparison of the derived FIR properties (FIR luminosity, L FIR , dust mass, M dust , dust temperature, T, emissivity index, β) with more metal-rich galaxies and investigate the detection of a potential submm excess. Methods. The data reduction method is adapted for each galaxy in order to derive the most reliable photometry from the final maps. The derived PACS flux densities are compared with the Spitzer MIPS 70 and 160 µm bands. We use colour−colour diagrams to analyse the FIR/submm behaviour of the DGS galaxies and modified blackbody fitting procedures to determine their dust properties. To study the variation in these dust properties with metallicity, we also include galaxies from the Herschel KINGFISH sample, which contains more metal-rich environments, totalling 109 galaxies. Results. The location of the DGS galaxies on Herschel colour−colour diagrams highlights the differences in dust grain properties and/or global environments of low-metallicity dwarf galaxies. The dust in DGS galaxies is generally warmer than in KINGFISH galaxies (T DGS ∼ 32 K and T KINGFISH ∼ 23 K). The emissivity index, β, is ∼1.7 in the DGS, however metallicity does not make a strong effect on β. The proportion of dust mass relative to stellar mass is lower in low-metallicity galaxies: M dust /M star ∼ 0.02% for the DGS versus 0.1% for KINGFISH. However, per unit dust mass, dwarf galaxies emit about six times more in the FIR/submm than higher metallicity galaxies. Out of the 22 DGS galaxies detected at 500 µm, about 41% present an excess in the submm beyond the explanation of our dust SED model, and this excess can go up to 150% above the prediction from the model. The excess mainly appears in lower metallicity galaxies (12 + log(O/H) 8.3), and the strongest excesses are detected in the most metal-poor galaxies. However, we also stress the need for observations longwards of the Herschel wavelengths to detect any submm excess appearing beyond 500 µm.

Aims. In this work, we aim to provide a consistent analysis of the dust properties from metal-poor to metal-rich environments by linking them to fundamental galactic parameters. Methods. We consider two samples of galaxies: the Dwarf Galaxy Survey (DGS) and the Key Insights on Nearby Galaxies: a FarInfrared Survey with Herschel (KINGFISH), totalling 109 galaxies, spanning almost 2 dex in metallicity. We collect infrared (IR) to submillimetre (submm) data for both samples and present the complete data set for the DGS sample. We model the observed spectral energy distributions (SED) with a physically-motivated dust model to access the dust properties: dust mass, total-IR luminosity, polycyclic aromatic hydrocarbon (PAH) mass fraction, dust temperature distribution, and dust-to-stellar mass ratio. Results. Using a different SED model (modified black body), different dust composition (amorphous carbon in lieu of graphite), or a different wavelength coverage at submm wavelengths results in differences in the dust mass estimate of a factor two to three, showing that this parameter is subject to non-negligible systematic modelling uncertainties. We find half as much dust with the amorphous carbon dust composition. For eight galaxies in our sample, we find a rather small excess at 500 µm (≤1.5σ). We find that the dust SED of low-metallicity galaxies is broader and peaks at shorter wavelengths compared to more metal-rich systems, a sign of a clumpier medium in dwarf galaxies. The PAH mass fraction and dust temperature distribution are found to be driven mostly by the specific star formation rate, sSFR, with secondary effects from metallicity. The correlations between metallicity and dust mass or total-IR luminosity are direct consequences of the stellar mass-metallicity relation. The dust-to-stellar mass ratios of metal-rich sources follow the well-studied trend of decreasing ratio for decreasing sSFR. The relation is more complex for low-metallicity galaxies with high sSFR, and depends on the chemical evolutionary stage of the source (i.e. gas-to-dust mass ratio). Dust growth processes in the ISM play a key role in the dust mass build-up with respect to the stellar content at high sSFR and low metallicity. Conclusions. We conclude that the evolution of the dust properties from metal-poor to metal-rich galaxies derives from a complex interplay between star formation activity, stellar mass, and metallicity.