The MOSDEF Survey: Electron Density and Ionization Parameter at $z\sim2.3$

TL;DR

This study uses MOSDEF rest-frame optical spectroscopy to measure electron densities and the ionization state of star-forming regions at $z\sim2.3$, employing [O II] and [S II] doublets to derive $n_e$ and analyzing $O_{32}$, $R_{23}$, and $O3N2$ to infer the ionization parameter $\mathcal{U}$. A key finding is that $n_e$ is about $2.5\times10^{2}\ \mathrm{cm}^{-3}$ at $z\sim2.3$, roughly tenfold higher than local values, yet at fixed metallicity the ionization parameter appears similar to local galaxies, suggesting the primary redshift evolution in line ratios arises from metallicity evolution rather than changes in $\mathcal{U}$ or the hardening of the ionizing spectrum. The work shows that high-$z$ galaxies occupy the same $O_{32}$–$R_{23}$ and $O_{32}$–O3N2 sequences as the local, low-metallicity tail, while nitrogen-based indicators reveal an elevated N/O at fixed O/H driving the offset in the [N II] BPT diagram. These results imply that local gas-phase and stellar metallicities largely set the ionization state across cosmic time, with the notable caveats of nitrogen enrichment and diffuse-ionized gas effects affecting interpretation of certain line ratios.

Abstract

Using observations from the MOSFIRE Deep Evolution Field (MOSDEF) survey, we investigate the physical conditions of star-forming regions in $z\sim2.3$ galaxies, specifically the electron density and ionization state. From measurements of the [O II]$λλ$3726,3729 and [S II]$λλ$6716,6731 doublets, we find a median electron density of $\sim250$ cm$^{-3}$ at $z\sim2.3$, an increase of an order of magnitude compared to measurements of galaxies at $z\sim0$. While $z\sim2.3$ galaxies are offset towards significantly higher O$_{32}$ values relative to local galaxies at fixed stellar mass, we find that the high-redshift sample follows a similar distribution to the low-metallicity tail of the local distribution in the O$_{32}$ vs. R$_{23}$ and O3N2 diagrams. Based on these results, we propose that $z\sim2.3$ star-forming galaxies have the same ionization parameter as local galaxies at fixed metallicity. In combination with simple photoionization models, the position of local and $z\sim2.3$ galaxies in excitation diagrams suggests that there is no significant change in the hardness of the ionizing spectrum at fixed metallicity from $z\sim0$ to $z\sim2.3$. We find that $z\sim2.3$ galaxies show no offset compared to low-metallicity local galaxies in emission line ratio diagrams involving only lines of hydrogen, oxygen, and sulfur, but show a systematic offset in diagrams involving [N II]$λ$6584. We conclude that the offset of $z\sim2.3$ galaxies from the local star-forming sequence in the [N II] BPT diagram is primarily driven by elevated N/O at fixed O/H compared to local galaxies. These results suggest that the local gas-phase and stellar metallicity sets the ionization state of star-forming regions at $z\sim0$ and $z\sim2$.

Figures (14)

• Figure 1: $R$ vs. $n_e$ curves (equation \ref{['eq:frat']}) from IRAF temden (red) and our five-level atom python script using new atomic data (black), where $R=$[O ii]$\lambda$3729/$\lambda$3726 (solid) or [S ii]$\lambda$6716/$\lambda$6731 (dashed).
• Figure 2: [O ii]$\lambda\lambda$3726,3729 doublets (left column) and [S ii]$\lambda\lambda$6716,6731 doublets (right column) from six different objects over a range of line ratios and densities. The black line shows the continuum subtracted spectrum for each object. The light gray band indicates the error spectrum for each object, while the blue and red lines show the Gaussian profile fits to the blueward and redward component of each doublet, respectively. The green line shows the total [O ii] profile. In each panel, the spectrum has been normalized so that the blue component has a peak height of unity. The line ratio $R$, density ($n_e$) in cm$^{-3}$, and redshift is displayed for each object, with $R=$[O ii]$\lambda$3729/$\lambda$3726 in the left column and $R=$[S ii]$\lambda$6716/$\lambda$6731 in the right column.
• Figure 3: Comparison of density estimates from the [O ii] and [S ii] doublets for a sample of local HII regions with high-S/N, high-resolution spectra. Black points denote density measurements for individual HII regions. The four red squares show limits plotted at the upper 1$\sigma$ uncertainty bound on the [S ii] density for objects that have higher [S ii]$\lambda$6716/$\lambda$6731 than the maximum theoretically allowed value. The black dashed line shows a one-to-one relationship. The blue line and shaded blue region show the best-fit line and 1$\sigma$ confidence interval, respectively. Parameters of the best-fit line are shown in equation \ref{['eq:bestfit']}.
• Figure 4: [O ii]$\lambda3729/\lambda3726$ (top) and [S ii]$\lambda6716/\lambda6731$ (bottom) line ratio distributions for 43 and 26 $z\sim2.3$ star-forming galaxies, respectively. In each panel, the dotted black line shows the median line ratio (corresponding to an electron density of 225 cm$^{-3}$ for [O ii] and 290 cm$^{-3}$ for [S ii]), while the dashed red lines show the minimum and maximum theoretically-allowed line ratios from Table \ref{['tab:bestfit']}.
• Figure 5: [S ii]$\lambda6716/\lambda6731$ line ratio distribution for local star-forming galaxies from SDSS, with lines as in Figure \ref{['fig:mosdefratiodist']}. The median line ratio for local galaxies (corresponding to an electron density of 26 cm$^{-3}$) falls near the low-density limit.