M^3D: Mosaicking M33 with MUSE datacubes I. Unveiling the diversity of HII regions in M33 with MUSE

TL;DR

This work delivers a deep, parsec-scale census of ionised gas in M33 by mapping a $3\times8$ arcmin$^2$ MUSE mosaic at $\approx 5$ pc resolution. Through a two-stage spectral fitting pipeline and robust segmentation, the authors identify 131 H II regions and characterize their emission, attenuation, and metallicity properties, revealing a wide diversity in ionisation states and nebular geometries, including both density-bounded and ionisation-bounded regions. The analysis combines spatially resolved line-ratio maps with radial profiles, showing how ionising sources (including WR stars and embedded clusters) and nebular geometry drive leakage of ionising photons and the hardness of the radiation field. The study also demonstrates the power of multi-band data (including JWST/MIRI and ancillary HST/WFC3 data) to connect small-scale feedback processes to the larger-scale ISM environment, setting the stage for advanced multi-cloud photoionisation modelling to interpret parsec-scale structure in galaxies.

Abstract

We present new VLT/MUSE observations of a 3 $\times$ 8 arcmin$^2$ mosaic along the southern major axis of the Local Group galaxy M33. These data provide an unprecedented view of the galaxy's interstellar medium (ISM) and allow us to resolve ionised nebulae at a spatial scale of $\approx$5 pc. We identified and catalogued 131 HII regions, down to Hαluminosities of $\approx 5\times$10$^{35}$ erg s$^{-1}$, one order of magnitude fainter than previous surveys on nearby galaxies beyond the Local Group, and we compared these regions with the spatial distribution of ionising stars and embedded star clusters. For each region, we extracted the corresponding integrated optical spectra and measured the intensity of key optical emission lines (Hβ, [OIII], [NII], Hα, [SII], [SIII]), other weaker optical lines when detectable, and Paschen lines to characterise the physical properties of the ioinised gas, such as density, dust attenuation, and metallicity. Our spatially resolved line ratio and flux maps reveal a remarkable diversity in ionisation properties, from dust-obscured regions hosting young stellar objects to highly ionised bubbles exhibiting high [OIII]/Hβratios. Our data reveal a diversity of ionisation fronts, ranging from well-defined to partial to absent. The radial profiles we obtained indicate the presence of both optically thin (density-bounded) HII regions permitting the escape of ionising photons and fully ionised, optically thick (ionisation-bounded) HII regions. The richness of this MUSE mosaic offers an unprecedented view of the ionised ISM at $\approx$5 pc resolution, providing direct insight into how stellar feedback shapes its environment.

Figures (16)

• Figure 1: Image of M33 using $g-r-i$ SDSS bands and showing the area of the galaxy targeted by MUSE (white contours). The positions of the centres of the individual MUSE pointings (1-24) are also shown. The red-coloured contours correspond to areas of high H$\alpha$ emission from the ground-based image of Massey2006.
• Figure 2: Emission maps of M33 MUSE data. Left: three-colour image from MUSE mapping the ionised gas emission: [Oiii]$\lambda5007$ in blue, H$\alpha$ in green and [Sii]$\lambda6716,6731$ in red. Light-blue stars are the 552 stars selected from the PHATTER survey Williams2021 as described in Appendix \ref{['sec:phatter']}. Smaller and larger star symbols denote two ranges of F475W magnitude, between 20 and 21 and below 20, respectively. Orange stars are WR stars from Neugent2011Neugent2019 (Appendix \ref{['sec:wr']}). White solid and dashed circles denote SNRs from Lee2014 and PNe from Ciardullo2004, respectively (Appendix \ref{['sec:snr']} and \ref{['sec:pne']}). Right: H$\alpha$ map from MUSE with the final configuration of the nebular masks, colour-coded on the basis of their classification as labelled in the legend.
• Figure 3: Optical [Oiii]/H$\beta$ versus [Nii]/H$\alpha$ (top) and [Sii]/H$\alpha$ (bottom) diagnostic diagrams from Baldwin1981 and Veilleux1987, respectively. Continuous and dashed curves show the demarcation criteria by Kewley2001Kewley2006 and Kauffmann2003, respectively. The dotted line indicates the traditional [Sii]/H$\alpha$ cutoff of 0.4 for SNR Long2018. Green stars are the integrated measurements of Hii regions, blue circles of PNe and red squares of SNRs as labelled in the legend. The Hii regions with yellow edges host WR stars. Black dots indicate the position in the diagram of three Hii regions (IDs 112, 113 and 120) which are discussed in more details in Sect. \ref{['sec:neb_geo']}.
• Figure 4: Metallicity gradients inferred from the Pilyugin2016 S-calibration (green) and the N$_2$S$_2$H$\alpha$ (violet) from Brazzini2024. The dashed green and violet lines indicate the best-fits to our data, and the shaded area indicates the 1-$\sigma$ scatter. Blue and gray lines and shaded areas are the gradients obtained from direct oxygen abundance measurements by Magrini2010 and Rogers2022, respectively. The stars yellow-edged stars are Hii regions hosting WR stars.
• Figure 5: Line-ratio maps, corrected for dust attenuation, of our M33 data for a S/N of the lines of interest larger than 2.5. From left to right: [Sii]$\lambda6716,6731$/H$\alpha$, [Siii]$\lambda9069$/[Sii]$\lambda6716,6731$, [Oiii]$\lambda5007$/[Sii]$\lambda6716,6731$, and [Oiii]$\lambda5007$/H$\alpha$. Black contours indicate the dendrograms classified as Hii regions. In the [Sii]/H$\alpha$ maps (panel a), the purple dashed contours show the level value of 0.35.