Table of Contents
Fetching ...

Excitation and Ionization Properties of Star-forming Galaxies at z=2.0-9.3 with JWST/NIRSpec

Ryan L. Sanders, Alice E. Shapley, Michael W. Topping, Naveen A. Reddy, Gabriel B. Brammer

TL;DR

This study leverages JWST/NIRSpec spectroscopy of 164 star-forming galaxies across z=2.0–9.3 to map excitation and ionization conditions using rest-frame optical line diagnostics. By constructing composite spectra in five redshift bins and comparing to local benchmarks, the authors find a common high-excitation sequence in the BPT diagrams and an O32–R23 locus that remains largely unchanged from z~2 to z~6, implying that ionization conditions do not strongly evolve over 2.4 Gyr of cosmic history. The inferred metallicities are typically ~0.1–0.3 $Z_{igodot}$ with uniformly hard ionizing spectra, consistent with alpha-enhancement and younger stellar populations at high redshift. The results support the applicability of local diagnostic calibrations to high-z samples within this range and highlight the need for larger, deeper samples to refine the high-redshift ionization landscape and potential LyC-leakage indicators.

Abstract

We utilize medium-resolution JWST/NIRSpec observations of 164 galaxies at $z=2.0-9.3$ from the Cosmic Evolution Early Release Science (CEERS) survey to investigate the evolution of the excitation and ionization properties of galaxies at high redshifts. Our results represent the first statistical constraints on the evolution of the [OIII]/H$β$ vs. [NII]/H$α$, [SII]/H$α$, and [OI]/H$α$ ``BPT'' diagrams at $z>2.7$, and the first analysis of the O32 vs. R23 diagram at $z>4$ with a large sample. We divide the sample into five redshift bins containing 30-40 galaxies each. The subsamples at $z\sim2.3$, $z\sim3.3$, and $z\sim4.5$ are representative of the main-sequence star-forming galaxy population at these redshifts, while the $z\sim5.6$ and $z\sim7.5$ samples are likely biased toward high specific star-formation rate due to selection effects. Using composite spectra, we find that each subsample at $z=2.0-6.5$ falls on the same excitation sequence in the [NII] and [SII] BPT diagrams and the O32-R23 diagram on average, offset from the sequences followed by $z=0$ HII regions in the same diagrams. The direction of these offsets are consistent with high-redshift star-forming galaxies uniformly having harder ionizing spectra than typical local galaxies at fixed nebular metallicity. The similarity of the average line ratios suggests that the ionization conditions of the interstellar medium do not strongly evolve between $z\sim2$ and $z\sim6$. Overall, the rest-optical line ratios suggest the $z=2.7-9.3$ CEERS/NIRSpec galaxies at log($M_*/M_{\odot})\sim7.5-10$ have high degrees of ionization and moderately low oxygen abundances ($\sim0.1-0.3~Z_{\odot}$), but are not extremely metal poor ($<0.1~Z_{\odot}$) even at $z>6.5$.

Excitation and Ionization Properties of Star-forming Galaxies at z=2.0-9.3 with JWST/NIRSpec

TL;DR

This study leverages JWST/NIRSpec spectroscopy of 164 star-forming galaxies across z=2.0–9.3 to map excitation and ionization conditions using rest-frame optical line diagnostics. By constructing composite spectra in five redshift bins and comparing to local benchmarks, the authors find a common high-excitation sequence in the BPT diagrams and an O32–R23 locus that remains largely unchanged from z~2 to z~6, implying that ionization conditions do not strongly evolve over 2.4 Gyr of cosmic history. The inferred metallicities are typically ~0.1–0.3 with uniformly hard ionizing spectra, consistent with alpha-enhancement and younger stellar populations at high redshift. The results support the applicability of local diagnostic calibrations to high-z samples within this range and highlight the need for larger, deeper samples to refine the high-redshift ionization landscape and potential LyC-leakage indicators.

Abstract

We utilize medium-resolution JWST/NIRSpec observations of 164 galaxies at from the Cosmic Evolution Early Release Science (CEERS) survey to investigate the evolution of the excitation and ionization properties of galaxies at high redshifts. Our results represent the first statistical constraints on the evolution of the [OIII]/H vs. [NII]/H, [SII]/H, and [OI]/H ``BPT'' diagrams at , and the first analysis of the O32 vs. R23 diagram at with a large sample. We divide the sample into five redshift bins containing 30-40 galaxies each. The subsamples at , , and are representative of the main-sequence star-forming galaxy population at these redshifts, while the and samples are likely biased toward high specific star-formation rate due to selection effects. Using composite spectra, we find that each subsample at falls on the same excitation sequence in the [NII] and [SII] BPT diagrams and the O32-R23 diagram on average, offset from the sequences followed by HII regions in the same diagrams. The direction of these offsets are consistent with high-redshift star-forming galaxies uniformly having harder ionizing spectra than typical local galaxies at fixed nebular metallicity. The similarity of the average line ratios suggests that the ionization conditions of the interstellar medium do not strongly evolve between and . Overall, the rest-optical line ratios suggest the CEERS/NIRSpec galaxies at log( have high degrees of ionization and moderately low oxygen abundances (), but are not extremely metal poor () even at .
Paper Structure (14 sections, 6 equations, 6 figures, 1 table)

This paper contains 14 sections, 6 equations, 6 figures, 1 table.

Figures (6)

  • Figure 1: Composite spectra of samples in each redshift interval, constructed as described in Sec. \ref{['sec:stacks']}. Each spectrum has been normalized to the measured H$\beta$ intensity and offset vertically for display purposes. The inset panel zooms in on the the weak red rest-optical lines for the $z<6.5$ composites on the same scale, vertically offset for clarity. The measured line ratios and uncertainties are reported in Table \ref{['tab:stacks']}.
  • Figure 2: SFR vs. M$_*$ for samples at $z=2.0-6.5$. Filled circles with error bars display SFR(H$\alpha$), while hollow squares present SFR(SED). In each panel, the dashed line displays the star-forming main sequence from the parameterization of spe2014, calculated at the median redshift of the galaxies in each bin. The gray two-dimensional histogram displays the distribution of $z\sim0$ star-forming galaxies from SDSS, for which the orange line shows the median trend.
  • Figure 3: The [N ii] (left column), [S ii] (middle column), and [O i] (right column) BPT diagrams at $z=2.0-6.5$. Each row displays individual galaxies and composite spectra for a single redshift interval. Arrows denote 3$\sigma$ upper limits when the weakest line was not detected. The gray two-dimensional histogram shows the distribution of $z\sim0$ star-forming galaxies from SDSS, while the orange line provides the median trend of this sample. The black solid line displays the local empirical demarcation between star-forming galaxies and AGN from kau2003 ([N ii] BPT) or kew2006 ([S ii] and [O i] BPT diagrams).
  • Figure 4: The [N ii] (left), [S ii] (middle), and [O i] (right) BPT diagrams for samples at $z=0-6.5$. Composite spectra and individually-detected CEERS galaxies are displayed as in Fig. \ref{['fig:bptz']}. Composite spectra of star-forming galaxies at $z\sim1.5$ and $z\sim2.3$ from the MOSDEF survey are displayed as "X"s sha2019. The distribution of $z\sim0$ star-forming galaxies from SDSS is shown in the gray two-dimensional histogram, with the median trend provided in orange. The brown dashed line denotes the median line ratio sequences derived from the catalog of $\sim1000$$z=0$ H ii regions presented in san2017pil2016.
  • Figure 5: The O$_{32}$ vs. R$_{23}$ diagram at $z=2.0-9.3$. Both line ratios have been corrected for dust reddening. Each row displays individual galaxies and composite spectra for a single redshift interval. The gray two-dimensional histogram and orange line shows $z\sim0$ star-forming galaxies as in previous figures.
  • ...and 1 more figures