DREAMS.II. Galaxy Demographics from Direct Te-Based Metallicities at z~2-10: Tracing the Evolution of the Mass-Metallicity and Fundamental Relations

TL;DR

This study uses JWST/NIRSpec spectra from DREAMS, JADES, ERO, and CEERS to derive direct $T_e$-based metallicities for 292 galaxies in the range $z o 2$–10, employing spectral stacking to detect the faint [O III] λ4363 auroral line. By analyzing strong-line diagnostics, photoionization models, and the dependence on stellar mass and SFR, the authors map the evolution of the MZR and the FMR, finding a monotonic metallicity decline at fixed $M_*$ from $z o 3$ to $10$, and a clear FMR offset at $z o 8$–10. The results show that high-$z$ ISM conditions lie at the low-metallicity, high-ionization end of the local $U$–$Z$ anti-correlation, with photoionization models reproducing the observed line-ratio evolution. A comparison with strong-line calibrations underscores deficiencies of local calibrations at high redshift, highlighting the importance of direct Te metallicities. Using the ChemicalUniverseMachine, the paper argues that the FMR break can be explained by enhanced gas inflows or outflows (or both), implying a non-equilibrium baryon cycle in the early universe with substantial gas fractions and diluted metal enrichment, and offering constraints on galaxy growth during the first billion years.

Abstract

We present the statistics of line ratios and direct Te-based metallicities from JWST medium-resolution spectra of 292 galaxies at z=2-10, combining DREAMS observations with those of JADES and CEERS. To remove systematics caused by stellar mass (M*) and star formation rate (SFR), we construct stacked spectra binned by redshift within fixed M* and SFR ranges, as well as across the full ranges. We find that the [OII]3727/Hb ratio drops by a factor of five from z~3 to 8 at fixed M* and SFR, in contrast to the nearly constant [OIII]5007/Hb ratio. We derive metallicities via the direct Te method using the [OIII]4363 line, and identify that high-z galaxies lie on the low-metallicity end of the anti-correlation between ionization parameter and metallicity at z~0. Photoionization modeling demonstrates that the redshift evolution, where metallicity decreases and ionization parameter increases, self-consistently explains the observed line ratios. We then examine the mass-metallicity (MZ) and fundamental (MZ-SFR) relations. Including additional galaxies at z~10-12, we find that the MZ relation monotonically decreases from z~3 to 10 at fixed M*, while the MZ-SFR relation shows a significant decline at z>8. Based on the ChemicalUniverseMachine model, this evolutionary trend can be explained by enhanced gas inflow (outflow) by a factor of ~5 (~1.7) at z>8.

Figures (16)

• Figure 1: The BPT diagram ([O iii]$\lambda$5007/H$\beta$ vs. [N ii]$\lambda$6584/H$\alpha$) used for the classification of galaxies and the exclusion of AGN. We plot all galaxies from the parent catalog for which the relevant emission lines are covered and at least one line in each ratio is detected with S/N $>$ 3. Circles represent galaxies where all four lines are detected. Triangles indicate 3$\sigma$ upper or lower limits for galaxies where only one line in a given ratio is detected. The solid curves show the AGN/SFG demarcation lines of Kewley01 (upper) and Kauffmann03 (lower). Galaxies are color-coded according to their classification. Those identified as AGN based on these criteria are shown in red and are excluded from our final sample. Galaxies exhibiting broad H$\alpha$ emission are shown in blue and are excluded from our analysis as broad-line AGN, regardless of their position on the diagram. The remaining galaxies, classified as SFGs and used in our analysis, are shown in gray.
• Figure 2: Star formation rate versus stellar mass for the galaxy sample used in this study, shown in five redshift bins: $z = 2–4,\, 4–5.5,\, 5.5–7,\, 7–9.5,$ and $z=9.5-12$. Filled gray dots represent our sample galaxies, with stellar masses derived from Prospector SED fitting, and SFRs calculated from dust-corrected H$\beta$ luminosities. In the highest-redshift panel, open gray dots indicate literature data points. Note that the methods used to estimate stellar mass and SFR differ across these studies. The blue solid line in each panel shows the predicted star formation main sequence from the simulation by Popesso23 at the corresponding redshift. For comparison, we show samples used in previous FMR studies: the gray dotted line shows the local ($z \sim 0$) relation from AM13, while the gray dashed and dot-dashed lines correspond to $z \sim 4$–10 observations from Nakajima23 and Curti23, respectively. In the $z \sim 3$ panel, gray symbols represent the sample from Sanders21. Red rectangles, shown in all panels, denote the selection criteria for the subsample, which is designed to minimize the sample bias and effects of FMR definition by comparing galaxies within a consistent region of the SFR-$M_*$ plane.
• Figure 3: Stacked spectra used in this study, shown for four redshift bins ($z \sim 3$, 4, 6, and 8; from top to bottom) and three wavelength ranges (from left to right): [O ii]$\lambda$3727; H$\gamma$ and [O iii]$\lambda$4363; and H$\beta$, [O iii]$\lambda$4959, and [O iii]$\lambda$5007. Gray vertical lines mark the expected positions of the relevant emission lines. Black curves show the stacked spectra normalized by the H$\beta$ flux of each galaxy before stacking. Red curves indicate the Gaussian fits used for emission-line flux measurements.
• Figure 4: Same as Figure \ref{['fig:stack_spectra']}, but for sample in a fixed $M_*$-SFR subsample.
• Figure 5: Redshift evolution of strong-line indices measured from stacked spectra. The horizontal axis shows redshift ($z \sim 3$, 4, 6, and 8), and the vertical axis shows the logarithmic values of the line ratios: R3 ([O iii]$\lambda$5007/H$\beta$); blue triangles) and R2 ([O ii]$\lambda$3727/H$\beta$); green circles). Filled symbols represent measurements from the full sample at each redshift, while open symbols correspond to subsamples matched in stellar mass and SFR. The lines indicate predictions from Cloudy photoionization models, calculated using the observed gas-phase metallicity and ionization parameter ($\log U$) derived for each redshift bin from the full sample.