JWST Census for the Mass-Metallicity Star-Formation Relations at z=4-10 with the Self-Consistent Flux Calibration and the Proper Metallicity Calibrators

TL;DR

This study leverages JWST/NIRSpec data from the ERO, GLASS, and CEERS programs to chart the mass–metallicity relation and its dependence on star formation from z=4 to z=10. By instituting a self-consistent flux-calibration workflow and anchoring strong-line metallicity indicators to direct $T_e$ measurements from 10 galaxies with [O III] 4363 detections, the authors construct robust MZ and SFR–MZ relations across a wide redshift range. They find only modest evolution in the MZ relation from z≈2–3 to z=4–10, but the SFR–MZ relation shows no evolution up to z≈8 and a significant decrease beyond z>8, implying a potential break in metallicity equilibrium driven by gas inflows and outflows. The work provides a large, consistently reduced high-z metallicity dataset, offering critical constraints for galaxy chemical evolution models and informing interpretations of ionization conditions in the early universe.

Abstract

We present the evolution of the mass-metallicity (MZ) relations at z=4-10 derived with 135 galaxies identified in the JWST/NIRSpec data taken from the three major public spectroscopy programs of ERO, GLASS, and CEERS. Because there are many discrepancies between flux measurements reported by early ERO studies, we first establish our NIRSpec data reduction procedure for reliable emission-line flux measurements and errors successfully explaining Balmer decrements with no statistical tensions via thorough comparisons of the early ERO studies. Applying the reduction procedure to the 135 galaxies, we obtain emission-line fluxes for physical property measurements. We confirm that 10 out of the 135 galaxies with [OIII]4363-lines have electron temperatures of (1.1-2.3)*10^4K, similar to lower-z star-forming galaxies, that can be explained by heating of young massive stars. We derive metallicities of the 10 galaxies by the direct method and the rest of the galaxies with strong lines by the metallicity calibrations of Nakajima et al. (2022) applicable for these low-mass metal-poor galaxies, anchoring the metallicities with the direct-method measurements. We thus obtain MZ relations and star-formation rate (SFR)-MZ relations over z=4-10. We find that there is a small evolution of the MZ relation from z~2-3 to z=4-10, while interestingly that the SFR-MZ relation shows no evolution up to z~8 but a significant decrease at z>8 beyond the error. This SFR-MZ relation decrease at z>8 may suggest a break of the metallicity equilibrium state via star-formation, inflow, and outflow, while further statistical and local-baseline studies are needed for a conclusion.

Figures (5)

• Figure 1:
• Figure 2: Balmer emission line ratios of H$\gamma$/H$\beta$ and H$\delta$/H$\beta$ for the ERO objects (04590 in red, 05144 in emerald green, 06355 in orange, and 10612 in green) in the main panel. The colored filled circles represent the new emission line flux measurements and errors with our improved reduction, while the other symbols show the measurements based on the early data release, as shown in the legend curti2023_erorhoads2023_eroarellano-cordova2022_ero. For the measurements of arellano-cordova2022_ero, we adopt the average values based on the two spectra (o007 and o008) for 06355 and 10612, but refer only to the o008 value for 04590. For those with line ratios outside the ranges of the plot, we place them close to the edge of the plot with an arrow to indicate the direction toward the actually reported line ratio. The lines show expected Balmer decrements as a function of E(B$-$V) for the three different attenuation curves (cardelli1989 with dashed, calzetti2000 with dotted, and SMC (gordon2003) with dot-dashed) and the two different electron temperatures ($T_e=10000$ K in gray and $25000$ K in magenta), although the different attenuation curve and $T_e$ have a small impact on this diagram. The deviation of the H$\delta$/H$\beta$ ratio from the theoretically expected value based on the H$\gamma$/H$\beta$ ratio ($\Delta$ H$\delta$/H$\beta$ = ((H$\delta$/H$\beta$)${\rm obs}$$-$ (H$\delta$/H$\beta$)${\rm expected}$) normalized by (H$\delta$/H$\beta$)$_{\rm expected}$) is highlighted in the upper panel, and that of the H$\gamma$/H$\beta$ ratio based on the H$\delta$/H$\beta$ ratio in the left panel, for each symbol as in the main panel. The corresponding E(B$-$V) values are printed in these sub-panels, where the calzetti2000 attenuation curve and $T_e=17500$ K are adopted for reference. Our measurements all follow the sequence of the expected Balmer decrements within the uncertainties, while quite a few of the measurements from the earlier studies show the ratios that are not physically and/or consistently explainable.
• Figure 3: Redshift distribution of the high-redshift galaxy samples used in this work. The red histogram presents the JWST sample constructed in this study using our analyzed ERO, GLASS, and CEERS spectra. The blue histogram shows the samples compiled from the literature (see Section \ref{['sssec:results_metallicity_literature']}). The sum of these samples, represented by the gray histogram, is used to investigate the mass-metallicity relationships at $z=4-10$ in this work.
• Figure 4: O32 vs. R23 diagram comparing the JWST sample at $z=4-9$, analyzed in this work, with lower redshift samples at $z=2-3$. The JWST objects are shown as red filled symbols, with circles representing ERO, pentagons representing GLASS, and diamonds representing CEERS. Objects with metallicity measurements obtained using the direct $T_e$ method are indicated with red open circles. The gray open symbols represent the averages of lower-redshift samples of continuum-selected galaxies NO2014troncoso2014sanders2016onodera2016strom2017 and Ly$\alpha$ emitting galaxies (LAEs; NO2014erb2016nakajima2020), as indicated in the legend. Arrows indicate $3\sigma$ lower limits. Gray shading illustrates the equivalent distribution for nearby SDSS galaxies.
• Figure 5: The [N ii] BPT diagram. Red symbols represent the JWST objects, as shown in Figure \ref{['fig:o3o2r23']}, whose H$\alpha$$+$[N ii] lines are covered in the NIRSpec spectra. Arrows indicate $3\sigma$ limits. Two popular demarcation curves between AGNs and star-forming galaxies, from kewley2001 (dot-long dashed; 'Kew01') and kauffmann2003 (short dash-long dashed; 'Kau03'), are also shown. None of the objects fall above the demarcation curves significantly beyond the measurement uncertainties.