Unveiling Metal Mixing in a Grand-Design Spiral: A UV-optical multiphase spatially resolved study of M83

This paper presents a spatially resolved, multiphase study of chemical enrichment around young star clusters (YSCs) in the nearby grand-design spiral galaxy M83. The study combines far-ultraviolet (UV) absorption-line spectroscopy from HST/COS with co-spatial optical spectroscopy from VLT/MUSE and LBT/MODS. The sample includes 18 YSCs with spectroscopic ages of approximately 1–6 Myr and galactocentric radii out to R/R25 = 0.56.

The authors derive neutral (H I) abundances from UV absorption-line spectroscopy and compare them with ionised (H II) abundances from reddening-corrected optical emission lines. Because auroral lines are not detected in all regions, they develop and apply an empirical multi-zone electron-temperature (Te) calibration based on strong-line diagnostics to estimate Te and derive reliable nebular abundances.

The study measures oxygen (O), sulphur (S), nitrogen (N), and iron (Fe) abundances, tracing enrichment from distinct nucleosynthetic channels. The α-elements (O and S) exhibit similar behaviour, consistent with enrichment by core-collapse supernovae, whereas Fe shows weaker variations, reflecting its delayed production by Type Ia supernovae. Nitrogen displays the largest phase offset (ionized-neutral), with enhancements of up to ∆N/H∼1.5 dex and ∆N/O>1.5 dex in the ionised gas relative to the neutral phase, indicating localised enrichment by massive stars and inefficient mixing between gas phases on Myr timescales. While the ionised gas exhibits signatures of feedback-regulated chemical enrichment and large-scale abundance gradients, corresponding trends are weak or absent in the neutral gas, consistent with metals remaining largely confined to the immediate star-forming environment during the earliest stages of cluster evolution in a massive grand-design spiral.

Key findings from the paper include:

1. Element-dependent enrichment: Within the ionised phase, α-elements (O and S) behave coherently, consistent with their shared core-collapse supernova origin. Iron behaves differently compared to oxygen, with weaker correlation and large scatter, consistent with its delayed dominant Type Ia SNe contribution and its strong sensitivity to dust depletion. Nitrogen shows the most strikingly different behaviour: in M83 the ionised-neutral offsets reach ∆N/H ≈ 1.5 dex and persist over the several-Myr timescales probed here, pointing to localised enrichment into the ionised gas and inefficient short-timescale mixing of massive stellar wind-driven nitrogen yields into the cold neutral ISM. The magnitude and persistence of this offset in a massive spiral galaxy strongly contrast with NGC 5253 (dwarf galaxy), where nitrogen offsets diminish within ∼8 Myr, highlighting the role of galactic potential wells and stellar outflows in regulating chemicals.

2. Age dependence of enrichment clocks: The phase-offsets (ionised-neutral) of α-elements increase with cluster age, consistent with fresh CCSNe enrichment on ∼3-5 Myr timescales. Iron shows no clear age dependence in the ∼1-6 Myr window probed here, as expected from its delayed origin in Type Ia SNe, though depletion onto dust likely induces variation within these short timescales. Nitrogen, in contrast, remains elevated in the ionised phase across ∼1-6 Myr, with no sign of the rapid decline as seen in NGC 5253. Together, these trends imply that while enrichment occurs rapidly within the ionised phase, the transfer of metals into the neutral gas is likely to occur on longer timescales than the ∼1–6 Myr window probed here.

3. Outflows and environment-regulated mixing: The large, sustained ∆N/O (>1.5 dex) in M83 reflects slow cross-phase mixing in a deep potential well: M83’s stronger gravitational confinement likely confines feedback-driven outflows to the local environment, so wind momentum primarily reshapes and enriches the H II gas rather than dispersing metals into the larger H I reservoir within ≲6 Myr. This indicates a clear decoupling between ionised and neutral gas phases on Myr timescales, with enrichment remaining localised to the ionised medium. Consistent with this, excess nitrogen abundance in ionised gas, ∆N/H (=(N/H)HII - (N/H)HI), shows strong positive correlations with stellar feedback (luminosity and wind momentum), while other elemental excesses (∆X/H for X=O, S, Fe) show weak, insignificant trends.

4. Spatial distribution and mixing: Outside the nucleus (R/R25 ≳0.05), abundance trends in both the neutral and ionised gas are generally weak and statistically insignificant (p-value > 0.05). However, ionised gas exhibits anti-correlations with galactocentric distance for all elements, broadly consistent with previously reported radial metallicity gradients in M83, though these may flatten at larger radii or vary locally due to dynamical processes. Neutral gas abundances show mild positive trends with radius, potentially reflecting inside-out enrichment commonly seen in star-forming galaxies, but the low statistical significance prevents firm conclusions.

The paper concludes that while the fundamental nucleosynthetic clocks are universal (rapid nitrogen injection through massive stellar winds, prompt α-element production through CCSNe, and delayed iron enrichment through Type Ia SNe), their observed signatures between phases depend strongly on environment. In average-metallicity dwarfs, WR-driven yields are redistributed and mixed on relatively short timescales, producing measurable convergence between phases. In M83, by contrast, enrichment remains largely confined to the ionised phase, with large abundance offsets between H I and H II regions persisting over ∼6-8 Myr.

Improvements to AI Systems:

1. Multi-Phase Chemical Evolution Modeling: Enhance galaxy evolution simulations by incorporating phase-dependent (H I vs. H II) metal enrichment timescales, explicitly decoupling ionized-gas enrichment (prompt CCSNe and stellar winds) from neutral-gas enrichment (delayed mixing). This improves predictions of observed abundance gradients and scatter in massive spiral galaxies.

2. Environment-Aware Feedback Calibration: Update sub-grid feedback models in cosmological simulations to include gravitational potential well depth as a key regulator of metal mixing efficiency. The AI can now predict that in deep-potential galaxies (like M83), metals remain confined to H II regions for >6 Myr, whereas in shallow-potential dwarfs (like NGC 5253), mixing occurs within 8 Myr.

3. Element-Specific Enrichment Clocks: Train AI classifiers to use multi-element abundance ratios (e.g., N/O, O/Fe, S/O) as temporal diagnostics of star cluster age and nucleosynthetic phase, distinguishing between prompt CCSNe (α-elements), delayed Type Ia SNe (Fe), and wind-driven N injection. This enables age-dating of young clusters from chemical fingerprints alone.

4. Empirical Multi-Zone Temperature Calibration: Replace single-zone electron temperature assumptions with a machine-learned empirical calibration that predicts Te from strong-line diagnostics across multiple ionization zones. This improves abundance derivations in low-SNR spectra where auroral lines are undetectable, reducing systematic biases in metallicity estimates for distant galaxies.

5. Cross-Phase Mixing Timescale Predictor: Develop a predictive model that estimates the timescale for metal transfer from ionized to neutral gas phases based on galaxy mass, star formation rate, and gravitational potential. This AI can forecast whether a given galaxy will show large ionized-neutral abundance offsets (like M83) or rapid convergence (like NGC 5253).

6. Feedback-Metal Correlation Analyzer: Implement a regression tool that quantifies the correlation between stellar feedback metrics (luminosity, wind momentum) and elemental excesses (ΔN/H, ΔO/H) in ionized gas. This allows automated identification of which elements are most sensitive to local feedback, improving interpretations of observed abundance scatter in star-forming regions.

7. Radial Gradient Flattening Detector: Enhance spectral analysis pipelines to automatically detect and characterize radial abundance gradient flattening or local variations in both neutral and ionized phases, using statistical significance testing (p-value thresholds) to distinguish genuine gradients from stochastic local enrichment.

What the Improved AI System Can Do:

• Predict the spatial and temporal distribution of α-elements, N, and Fe in galaxies of varying mass, correctly reproducing the observed phase decoupling and enrichment confinement.

• Automatically classify young star clusters into age bins (1–3, 3–6 Myr) using multi-element abundance ratios, without relying on photometric or spectroscopic age indicators.

• Derive accurate nebular abundances from low-resolution spectra by applying the learned multi-zone Te calibration, extending metallicity studies to high-redshift galaxies.

• Forecast whether a galaxy will exhibit large ionized-neutral abundance offsets based on its gravitational potential and star formation history, aiding in the interpretation of future JWST and ALMA observations.

• Identify feedback-dominated regions in galaxies by correlating stellar wind momentum with nitrogen excess, enabling targeted follow-up of chemically enriched outflows.

• Provide robust radial abundance gradient maps that separate genuine large-scale trends from local enrichment noise, improving comparisons between simulations and observations.

We present a spatially resolved, multiphase study of chemical enrichment around young star clusters (YSCs) in the nearby grand-design spiral M83 by combining far-ultraviolet(UV) absorption-line spectroscopy from HST/COS with cospatial optical spectroscopy from VLT/MUSE and LBT/MODS. Our sample includes 18 YSCs spanning spectroscopic ages of 1-6 Myr and galactocentric radii out to R/R 25=0.56. Neutral (H I) abundances were derived from UV absorption-line spectroscopy and compared with ionised (H II) abundances from reddening-corrected optical emission lines. Because auroral lines are not detected in all regions, we develop and apply an empirical multi-zone electron temperature (T e) calibration based on strong-line diagnostics to estimate T e and derive reliable nebular abundances. We measure oxygen(O), sulphur(S), nitrogen(N), and iron(Fe) abundance tracing enrichment from distinct nucleosynthetic channels. The alpha-elements (O and S) exhibit similar behaviour, consistent with enrichment by core-collapse supernovae, whereas Fe shows weaker variations, reflecting its delayed production by Type Ia supernovae. Nitrogen displays the largest phase offset (ionised-neutral), with enhancements of up to Delta N/H 1.5 dex and Delta N/O>1.5 dex in the ionised gas relative to the neutral phase, indicating localised enrichment by massive stars and inefficient mixing between gas phases on Myr timescales. While the ionised gas exhibits signatures of feedback-regulated chemical enrichment and large-scale abundance gradients, corresponding trends are weak or absent in the neutral gas, consistent with metals remaining largely confined to the immediate star-forming environment during the earliest stages of cluster evolution in a massive grand-design spiral.