Resolved HII regions in NGC 253: Ionized gas structure and suggestions of a universal density-surface brightness relation

TL;DR

This study delivers a highly resolved census of ~2500 HII regions in NGC 253 using a full-disk VLT-MUSE mosaic, revealing that Hα emission profiles are best described by a two-component (double-Gaussian) structure, with a compact core (deconvolved HWHM ≈ 7 pc) surrounded by a diffuse halo (HWHM ≈ 30 pc) that contains about 80% of the flux. By inverting the emission profiles and using the [SII] diagnostic, the authors derive central electron densities in the range $n_e \,\sim\,10$–$100\ \mathrm{cm^{-3}}$, and show a strong correlation between central density and region surface brightness, a relation that persists across the PHANGS-MUSE+HST sample for nearby galaxies. The luminosity–radius analysis demonstrates a robust isophotal radius–luminosity relation, but with caveats about the physical meaning of the isophotal boundary; other radii show weak or debated correlations that depend on background subtraction and resolution. The results imply a dense, compact core feeding a more extended halo and suggest a potentially universal density–surface-brightness scaling for HII regions, with important implications for stellar feedback modeling and for interpreting unresolved regions in distant galaxies. The authors release a large region catalog and discuss limitations related to completeness, background treatment, and spatial resolution, highlighting the need to test these findings across more systems and with higher-resolution data.

Abstract

We use the full-disk VLT-MUSE mosaic of NGC 253 to identify 2492 HII regions and study their resolved structure. With an average physical resolution of 17 pc, this is one of the largest samples of highly resolved spectrally mapped extragalactic HII regions. Regions of all luminosities exhibit a characteristic emission profile described by a double Gaussian with a marginally resolved or unresolved core with radius <10 pc surrounded by a more extended halo of emission with radius 20-30 pc. Approximately 80% of the emission of a region originates from the halo component. As a result of this compact structure, the luminosity-radius relations for core and effective radii of HII regions depend sensitively on the adopted methodology. Only the isophotal radius yields a robust relationship in NGC 253, but this measurement has an ambiguous physical meaning. We invert the measured emission profiles to infer density profiles and find central densities of n_e = 10-100 cm-3. In the brightest regions, these agree well with densities inferred from the [SII]6716,30 doublet. The central density of HII regions correlates well with the surface brightness within the effective radius. We show that this same scaling relation applies to the recent MUSE+HST catalog for 19 nearby galaxies. We also discuss potential limitations, including completeness, impacts of background subtraction and spatial resolution, and the generality of our results when applied to other galaxies.

Figures (21)

• Figure 1: Top: MUSE extinction-corrected H$\alpha$ map of NGC 253 in grayscale congiu2025. Points show identified H$\alpha$ peaks colored by decile of background-subtracted H$\alpha$ peak intensity (Section \ref{['sec:intbin']}, Figure \ref{['fig:LvsI']}). The brighter H2 regions tend to be located in the inner spiral arms while the fainter regions tend to be located in the outer spiral arms or in the interarm region. The black box shows the galactic center excluded from our analysis. Bottom Left: Cutout of the H$\alpha$ intensity map before extinction-correction to illustrate peak identification for one complex region (white box in top panel). White contours show the 25, 50, 84, 90, 99, and 99.5% intensity levels. White points show the peaks included in the final catalog and the black points show peaks removed by intensity, background contrast, and line ratio diagnostic cuts (Appendix \ref{['ap:bpt']}). Bottom Right: Luminosity of H2 regions (gray points) as a function of galactocentric radius. Colored points represent the median and 16-84% range for regions binned by H$\alpha$ peak intensity. The brightest regions tend to lie in the inner galaxy, reflecting the difficulty of detecting faint H2 regions in complex galactic environments.
• Figure 2: Median integrated flux profile, i.e., flux enclosed within the radius shown on the $x$-axis, over the full sample of H2 regions for several treatments of the background. Green shows our fiducial treatment, with $R_{\text{lum}}^{\text{max}} = 100$ pc, black shows $R_{\text{lum}}^{\text{max}} = 200$ pc, and the gray dashed line shows the integrated flux profile with no background subtraction. Here $R_{\text{lum}}^{\text{max}}$ is the maximum radius allowed when determining the aperture used to integrate the flux, calculate the background, and measure sizes for each H2 region. The gray dashed line shows that with no background subtraction the flux (and so also $R_{\text{50}}$ , $R_{\text{90}}$ , and $R_{\text{mom}}$) rises monotonically with distance from the peak. Our fiducial background subtraction leads to a flat enclosed flux, on average, between $\sim 50{-}150$ pc. Using a larger maximum aperture of $R_{\text{lum}}^{\text{max}} = 200$ pc (black line) yields a profile that rises out to $200$ pc, though more slowly than the unsubtracted case. Given this behavior, both our inferred luminosity and sizes based on fractions of enclosed flux, e.g., $R_{\text{50}}$ and $R_{\text{90}}$, are sensitive to the definition of background. The light blue line shows the median aperture radius for the whole sample using our fiducial treatment.
• Figure 3: Luminosity and intensity distributions in the H2 region catalog. Top left: Distribution of extinction-corrected H$\alpha$ intensities at the peaks that define our regions. The color bar indicates the ranges spanned by the peak intensity bins (deciles) used in our analysis. Top right: Relationship between extinction-corrected, background-subtracted L$_{\rm H\alpha}$ and peak intensity for individual regions (gray points) and median values within intensity bins (squares). In NGC 253, peak intensity and L$_{\rm H\alpha}$ are tightly correlated, and the black line shows the best fit to the bins. The green line shows expectations for an unresolved point source with individual traces for each field. Bottom left: Distribution of extinction-corrected, background-subtracted H$\alpha$ luminosities (L$_{\rm H\alpha}$) for our regions. Colored lines indicate the median luminosity within each peak intensity bin. Bottom right: The H$\alpha$ luminosity function for our regions with (black) and without (gold) background subtraction. For comparison, gray lines show the luminosity functions for each PHANGS-MUSE galaxy groves2023. Blue and red lines show power laws that match the upper and lower luminosity ranges for NGC 253.
• Figure 4: Background-subtracted radial profiles of extinction-corrected H$\alpha$ intensity for H2 regions grouped by their peak H$\alpha$ intensity. The individual profiles are normalized by the median peak intensity within the bin. The profiles of individual regions are plotted in gray and the median (solid black line) profile for each bin is drawn on top with the 16th to 84th percentiles shaded. The average MUSE PSF is shown with a black dotted line for comparison, normalized to the faintest region in the bin. Five radius measurements--HWHM (pink), $R_{\text{mom}}$ (purple), $R_{\text{50}}$ (yellow), $R_{\text{90}}$ (green), and $R_{\text{iso}}$ (blue)--are represented with vertical lines. A horizontal line is plotted at the fiducial isophote used to measure $R_{\text{iso}}$ , $10^{38.8}$ erg s$^{-1}$ kpc$^{-2}$. The profile shapes appear consistent across peak intensity bins. 23 H2 regions in bin 10 have profile peaks that extend off the plot.
• Figure 5: Luminosity and size for H2 regions in NGC 253. Top: 16th, 50th, and 84th percentile density contours of extinction-corrected, background-subtracted H$\alpha$ luminosity as a function of several radius measurements (Section \ref{['sec:radius']}). The large points show the median values of regions binned by peak H$\alpha$ intensity (Section \ref{['sec:intbin']}). There is no strong trend between HWHM (or core) radius and luminosity (Section \ref{['sec:hwhm']}), while the isophotal radius (blue) shows a strong correlation with luminosity (Section \ref{['sec:isophotal']}). The second moment radius, $R_{\text{mom}}$, and effective radius, $R_{\text{50}}$ , also show large scatter and little clear correlation with radius, while $R_{\text{90}}$ shows more ambiguous results that reflect its sensitivity to the adopted integration aperture (Section \ref{['sec:effectiveradius']}) Bottom: Luminosity-radius relations for each metric showing individual data points and results for different binning approaches: peak intensity bins (large colored markers), bins of H$\alpha$ luminosity (black circles), and bins of size (gray circles).