The JWST EXCELS survey: Insights into the nature of quenching at cosmic noon

TL;DR

This study tackles how massive galaxies quench during cosmic noon (1<z<3) by reconstructing detailed star formation histories from JWST/NIRSpec EXCELS spectroscopy combined with deep multi-band photometry. Using Bagpipes, the authors model SFHs with burst+decay components, quantify burst strengths, and derive quenching timescales, finding bursts of $\sim$100–1000 $M_\odot$ yr$^{-1}$ followed by rapid quenching, with a median post-starburst visibility of $\sim$ $612$ Myr and a broad range of quenching times $\tau_{q1}$ from $0.06$ to $1.75$ Gyr. They identify a significant, mass-dependent PSB population that could account for up to $\sim$73% of the growth of the massive quiescent population at $z\sim2$ and argue that SMGs are likely progenitors given peak SFRs and compact morphologies; they also find evidence for weak AGN activity in a subset of galaxies. The results imply multiple, fast quenching pathways after starbursts at cosmic noon, with AGN potentially sustaining quiescence but not necessarily initiating it, and highlight the role of compact, rapidly quenched systems in building the red sequence.

Abstract

We study 24 massive quiescent galaxies with $\log \textrm{M}_*/\textrm{M}_\odot > 10$ at $1 < z < 3$ with JWST/NIRSpec medium-resolution observations from the Early eXtragalactic Continuum and Emission Line Survey (EXCELS). We reconstruct their star formation histories and find that they have large bursts ($100\textrm{ M}_{\odot} \textrm{yr}^{-1} -1000 \textrm{ M}_{\odot} \textrm{yr}^{-1}$), followed by a rapid truncation of star formation. The number densities of the quenched galaxies in our sample that we predict underwent a submillimeter phase are consistent with submillimeter galaxies being the progenitors of our quenched population. The median post-starburst visibility time is $\sim600$ Myr, with more massive galaxies ($\log \textrm{M}_*/\textrm{M}_\odot > 10.7$) exhibiting shorter visibility times than lower mass galaxies. The range of quenching times -- defined as the time from the peak starburst to the time of quiescence -- found in this sample ($0.06-1.75$ Gyr) suggests multiple quenching pathways, consistent with previous studies. We do not see evidence for quenching mechanisms varying with redshift between $1<z<3$. We detect evidence for weak AGN activity in 4 out of the 8 galaxies with robust emission line detections, based on line ratio diagnostics. Our findings suggest that there are a diverse range of quenching mechanisms at cosmic noon, and support a scenario in which the primary quenching mechanisms are rapid ($<500$ Myr) following a starburst.

Figures (12)

• Figure 1: Super-color 1 (set by how red or blue the galaxy SED is) versus super-color 2 (determined by the strength of the 4000Å or Balmer break) based on photometry. Our sample of spectroscopically confirmed post-starburst and quiescent galaxies is shown in red and the underlying sample from the entire PRIMER field (de Lisle et al. in prep) is shown in gray. The dashed black lines indicate the boundaries between the post-starburst, quiescent, and star-forming regions.
• Figure 2: An example SFH (see equation \ref{['eq::sfh']}) used to fit our sample of galaxies. The galaxy in this model has a redshift of $z=2$, a burst age of $t_{\text{burst}} = 0.5$ Gyr, a formation age of $t_{\text{form}} = 3$ Gyr, and has formed $10^{11} M_\odot$ of stars. It has a burst mass fraction of 20 percent ($f_{\text{burst}} = 0.2$), an exponential decay timescale of $\tau_{\text{e}} = 3$ Gyr, and $\alpha = 50$ describing the declining timescale of the burst.
• Figure 3: Example Bagpipes fit of a post-starburst galaxy at $z=1.83$ (EXCELS -- 94982). Top: The input spectrum is shown in blue and the median of the posterior fit is shown in orange. The photometric measurements are shown as blue points and the black line denotes the fitted model without calibration corrections. The blue shading indicates regions masked during fitting. The lower panel shows the spectrophotometric calibration component. Bottom: The input photometry is shown as blue points, the posterior fitted photometry is shown in orange, and the orange line shows the posterior SED.
• Figure 4: Star formation rate as a function of galaxy stellar mass, where both properties have been derived from our full spectral fitting analysis described in Section \ref{['subsec:pipes_fitting']}. Triangles indicate values that fall below log(SFR)$<-2$. The dashed black line indicates where $\text{sSFR} < 0.2/t_{\text{universe}}$ at a redshift of $z=2$. The dashed red line indicates the star-forming main sequence at $z=2$ from Speagle2014.
• Figure 5: The star formation histories derived from Bagpipes for the galaxies in our sample. The labels on each plot indicate the EXCELS-ID followed by its super-color classification (Q, PSB, SF), as described in Section \ref{['sec:samp_sc']}. The black line indicates the median posterior in each age bin, the gray shading indicates the 68 percent confidence interval, and the light gray shading indicates the 95 percent confidence interval of the SFHs. The gray dashed lines show 10 random samples from the posterior distributions.