Table of Contents
Fetching ...

CLASSY XII: Nitrogen Enrichment Shaped by Gas Density and Feedback

Karla Z. Arellano-Córdova, Danielle A. Berg, Matilde Mingozzi, Bethan L. James, Fiorenzo Vincenzo, Noah S. J. Rogers, Evan D. Skillman, Ricardo O. Amorín, Fergus Cullen, Sophia R. Flury, Valentina Abril-Melgarejo, John Chisholm, Timothy Heckman, Matthew J. Hayes, Svea Hernandez, Nimisha Kumari, Chiaki Kobayashi, Claus Leitherer, Crystal L. Martin, Zorayda Martinez, Themiya Nanayakkara, Kaelee S. Parker, Peter Senchyna, Claudia Scarlata, Mabel G. Stephenson, Aida Wofford, Xinfeng Xu, Peixin Zhu

TL;DR

The CLASSY study uses direct-$T_e$ abundances for 45 local star-forming galaxies to map the $N/O$–$O/H$ chemical evolution and identify drivers of scatter, including electron density, star-formation activity, and feedback. It finds a plateau in $N/O$ at low metallicity, with substantial dispersion at higher $O/H$, and a robust link between $n_e$ and $N/O$, indicating density structure as a key scatter contributor. Comparison with high-$z$ galaxies shows no significant redshift evolution in $N/O$ at fixed metallicity up to $z\sim6$ (though $z>6$ UV-based nitrogen abundances can be offset), suggesting similar enrichment pathways across cosmic time. The work also demonstrates that feedback, compactness, and density together regulate nitrogen enrichment, and it presents chemical-evolution models in which wind retention and massive-star winds can explain the observed $N/O$ scatter, with implications for interpreting JWST-era observations of the early universe.

Abstract

We investigate the chemical evolution of N/O using a sample of 45 local star-forming galaxies (SFGs) from the CLASSY survey. This sample spans a wide range of galaxy properties, with robust determinations of nitrogen and oxygen abundances via the direct-$T_{\rm e}$ method. We explore how N/O relates to density structure, stellar mass, star formation rate (SFR), stellar age, compactness, and gas kinematics. In addition, we compare our results with those of galaxies at $z =2-10$ where N/O ratios were derived from optical or UV nitrogen lines, aiming to identify chemical enrichment pathways across cosmic time. Our analysis shows that the N/O-O/H relation in CLASSY galaxies aligns with the trends seen in local galaxies and extragalactic HII regions, and that galaxies at $z = 2-6$ exhibit similar N/O values, indicating no significant redshift evolution in N/O for a fixed metallicity. We identify a significant correlation between electron density $n_{\rm e}$([S II]) and N/O, suggesting that density structure contributes to the scatter in the N/O-O/H relation. The CLASSY galaxies with high SFRs or compact star formation show elevated N/O, though no strong correlation with stellar mass is found. We also find that high-velocity outflows (v$_{out}$ > 350 km/s) and low mass-loading factors are linked to elevated N/O, indicating that feedback plays a significant role. These results highlight the importance of density, star formation, and feedback from young stellar populations in shaping N/O enrichment and provide key insights for interpreting high-$z$ galaxies observed with JWST.

CLASSY XII: Nitrogen Enrichment Shaped by Gas Density and Feedback

TL;DR

The CLASSY study uses direct- abundances for 45 local star-forming galaxies to map the chemical evolution and identify drivers of scatter, including electron density, star-formation activity, and feedback. It finds a plateau in at low metallicity, with substantial dispersion at higher , and a robust link between and , indicating density structure as a key scatter contributor. Comparison with high- galaxies shows no significant redshift evolution in at fixed metallicity up to (though UV-based nitrogen abundances can be offset), suggesting similar enrichment pathways across cosmic time. The work also demonstrates that feedback, compactness, and density together regulate nitrogen enrichment, and it presents chemical-evolution models in which wind retention and massive-star winds can explain the observed scatter, with implications for interpreting JWST-era observations of the early universe.

Abstract

We investigate the chemical evolution of N/O using a sample of 45 local star-forming galaxies (SFGs) from the CLASSY survey. This sample spans a wide range of galaxy properties, with robust determinations of nitrogen and oxygen abundances via the direct- method. We explore how N/O relates to density structure, stellar mass, star formation rate (SFR), stellar age, compactness, and gas kinematics. In addition, we compare our results with those of galaxies at where N/O ratios were derived from optical or UV nitrogen lines, aiming to identify chemical enrichment pathways across cosmic time. Our analysis shows that the N/O-O/H relation in CLASSY galaxies aligns with the trends seen in local galaxies and extragalactic HII regions, and that galaxies at exhibit similar N/O values, indicating no significant redshift evolution in N/O for a fixed metallicity. We identify a significant correlation between electron density ([S II]) and N/O, suggesting that density structure contributes to the scatter in the N/O-O/H relation. The CLASSY galaxies with high SFRs or compact star formation show elevated N/O, though no strong correlation with stellar mass is found. We also find that high-velocity outflows (v > 350 km/s) and low mass-loading factors are linked to elevated N/O, indicating that feedback plays a significant role. These results highlight the importance of density, star formation, and feedback from young stellar populations in shaping N/O enrichment and provide key insights for interpreting high- galaxies observed with JWST.

Paper Structure

This paper contains 28 sections, 2 equations, 12 figures, 2 tables.

Figures (12)

  • Figure 1: Comparison between log(N/O) derived using $T_{\rm e}$[Oii] and log(N/O) derived using $T_{\rm e}$[Nii], labeled as TO2 and TN2, respectively. The solid line represents the 1:1 relation, and the results are color-coded with log($n_{\rm e}$[Sii]). Note that the uncertainties in the measurements of log(N/O) are larger when using $T_{\rm e}$[Nii] compared to $T_{\rm e}$[Oii]. It shows that the two values are consistent within the uncertainties, with a slight offset toward lower values of log(N/O) at low $n_{\rm e}$ when using $T_{\rm e}$[Nii].
  • Figure 2: The N/O-O/H relation for the CLASSY sample. CLASSY galaxies hosting WR stars are also identified (star symbols), and they show similar N/O ratios to the bulk of the sample, with the exception of the WR galaxy Mrk 996 james09, which exhibits a N/O enhancement. The comparison sample of local SFGs and Hii regions (grey symbols) was compiled and re-analyzed in arellanocordova24. A sample of high-redshift galaxies is included for comparison. SFGs at $z = 2-6$ with N/O ratios derived from optical [Nii] lines are shown as pentagons sanders23welch25rogers24arellanocordova25scholte25Zhang25, while squares represent $z > 6$ galaxies with N/O derived from UV nitrogen lines sanders23bmarqueschaves24jones23arellanocordova24schaerer24isobe23anakajima23topping24b. The different curves show predictions from chemical evolution models with an infall mass of log($M_{\rm inf}) = 10$ M$_\odot$, star formation efficiencies (SFE) of 0.5, 1, and 5 Gyr$^{-1}$, and a time infall of $\tau_{\rm inf} = 0.1$ Gyr represented by solid, dashed, and dotted lines, respectively vincenzo16. The CLASSY galaxies (orange circles) follow the expected N/O-O/H relation and occupy a similar parameter space as SFGs at $z=2-6$. In contrast, they differ from the $z>6$ population, showing no apparent redshift evolution.
  • Figure 3: The N/H-O/H relation for CLASSY and the sample of high redshift galaxies. SFGs at $z=2-6$ from sanders23bwelch25rogers24arellanocordova25scholte25 and Zhang25 in cyan pentagons, while the green squares represent $z>6$ galaxies from marqueschaves24topping24topping24bcurti24bschaerer24. For comparison, we include the local sample in gray circles. High redshift galaxies are shifted to higher values of N/H, mainly those galaxies using UV nitrogen lines (squares), while the optical N sample at $z =2-6$ follows the trend of CLASSY at $z\sim0$. The purple triangles show those CLASSY galaxies with N/O derived using UV [Niv] or [Niii] lines (see Martinez et al. 2025). The different curves show the predictions of chemical evolution models for the N/H–O/H relation, as in Fig. \ref{['fig:NO_highz']}, while the dash-dotted line represents an empirical calibration to SDSS galaxies from flury20.
  • Figure 4: Left N/O abundances as a function of low-ionization density, $n_{\rm e}$[Sii]. SFGs at $z>2$ are also included for comparison from sanders23welch25arellanocordova25stanton2025scholte25. We find that high density derived from $n_{\rm e}$[Sii] correlates with the high values of N/O for local galaxies (with a Kendall's coefficient of $\tau=0.372$ and p-value = 0.001), while $z=2-6$ galaxies show a similar trend. The yellow stars are galaxies with WR stars features. The best fit to the data for the N/O vs. log($n_{\rm e}$[Sii]) is shown in an orange solid line for CLASSY. Right: The N/O ratios as a function of high-ionization density, $n_{\rm e}$[Ariv]. N$^{+}$ and O$^{+}$ were derived using $n_{\rm e}$[Ariv] for 13 CLASSY galaxies. Only three CLASSY galaxies have measurements of both $n_{\rm e}$[Sii] and $n_{\rm e}$[Ariv] within a similar range of $n_{\rm e}$$=$ 200–600 cm$^{-3}$. J0808+3948 shows the highest values, with $n_{\rm e}$$=$ 1200–6310 cm$^{-3}$ and log(N/O) = –0.47 to –0.65 when derived using both density diagnostics.
  • Figure 5: Comparison of the N/O and N/H ratios as a function of the stellar mass (left), SFR (middle) and sSFR (right) for CLASSY. The circles indicate the sample of SFGs from berg12berg16berg19izotov17. The dashed line shows the N/O vs M$\star$ for SDSS galaxies derived by hayden-pawson22, while the dashed-dotted line represents the N/O vs. M$\star$ and N/H vs. M$\star$ relations from strom22 for galaxies at $z \sim 2-3$. Such relations were derived using strong-line methods. The dotted lines represents the solar value of N/O and N/H from asplund21. The pentagons and squares represent galaxies at $z =2-6$ and their N/O abundances on optical lines by sanders23welch25arellanocordova25scholte25Zhang25, while that the squares represent galaxies at $z >6$ based their N/O abundances on UV lines by isobe23amarqueschaves24topping24topping24bcurti24bschaerer24. Our best fit to the CLASSY data is shown in a solid blue line.
  • ...and 7 more figures