LEGA-C stellar populations scaling relations. I: Chemo-archaeological downsizing trends at z~0.7

TL;DR

This study analyzes 552 massive galaxies at 0.6 < z < 0.77 from the LEGA-C survey to derive light-weighted ages and metallicities via a Bayesian, index-plus-photometry approach (BaStA) with a comprehensive library of star-formation and metal-enrichment histories. It introduces revised absorption-index catalogs for LEGA-C DR3 and constructs robust, volume- and completeness-weighted scaling relations of age and metallicity with stellar mass and velocity dispersion, revealing downsizing and chemical downsizing already in place at z ~ 0.7. The results show a bimodal age distribution with a transition around M* ~ 1.2 × 10^11 M⊙ and a sharp age transition near log σ* ≈ 2.3, while the stellar metallicity–mass relation saturates above ~6 × 10^10 M⊙ and scales linearly with velocity dispersion, albeit with larger scatter at low dispersion. The work provides valuable constraints for galaxy evolution models and will enable cross-epoch comparisons, aided by publicly released index and stellar-population catalogs and cross-method comparisons with Prospector and Bagpipes.

Abstract

We analyze stellar population properties of 552 galaxies at redshift 0.6<z<0.77 from the LEGA-C spectroscopic survey. This first paper in a series presents the catalog of revised absorption indices for LEGA-C DR3 and inferred physical parameters, and derives benchmark scaling relations for the general massive galaxy population at intermediate redshift. We estimate light-weighted mean ages and stellar metallicities by interpreting key stellar absorption features and rizYJ photometry in a Bayesian framework with a comprehensive library of model spectra based on stochastic star formation and metallicity histories and dust attenuations. We discuss systematic uncertainties within our method and compared to other spectral fitting approaches. We derive volume-weighted scaling relations of light-weighted mean ages and stellar metallicities with stellar mass for the general galaxy population at <z>=0.7 and masses >10^10Msun. The downsizing trends observed locally were already in place 6 Gyr ago. We observe bimodal age distribution as a function of mass, transitioning around 10^11Msun. No bimodality appears in the stellar metallicity-mass relation, which changes from steep to flat across 10^10.8Msun. Similar trends emerge for age and metallicity with velocity dispersion, but with sharper transition from young to old around log(sigma)=2.3. Differences with respect to trens with stellar mass suggest that age primarily depends on velocity dispersion below and above the transition regime, while both stellar mass and velocity dispersion contribute to stellar metallicity. The catalogs of revised absorption index measurements for LEGA-C DR3 and inferred stellar population physical parameters will be released to public repositories. (Abridged)

Figures (18)

• Figure 1: Distribution in stellar mass and redshift for LEGA-C primary sample (grey dots and dot-dashed histograms, normalized to unit integral) and for the subsamples used in the analysis: silver sample (grey symbols and histograms); golden sample (golden symbols and histograms). The histograms for the silver and golden samples are normalized by the number of galaxies in the silver sample. The number of galaxies in each subsample is given in parenthesis. Redshifts come from the LEGA-C DR3 catalog, while stellar masses are those computed in this work.
• Figure 2: Left: Distribution in specific SFR versus stellar mass. Right: Distribution in U-V, V-J rest-frame colors. We separate galaxies into quiescent (Q, magenta circles) and star-forming (SF, blue stars) based on their location with respect to the Star Forming Main Sequence (solid line in the left panel): Q galaxies are those deviating by more than $2\sigma$ below the MS (i.e. they lie below the dashed line in the left panel). The right-hand panel illustrates how this selection compares with the UVJ quiescent zone outlined by the dashed segment Muzzin13b. Empty/filled symbols indicate galaxies in the silver/ golden samples. The numbers in the left panel give the number of SF and Q galaxies in the golden, silver and parent LEGA-C samples. All quantities in this plot come from the LEGA-C DR3 catalog, except stellar masses which are those computed in this work.
• Figure 3: Distribution in absorption indices diagnostic diagrams, showing the combinations of indices considered in this work for their different sensitivity to age and metallicity. The colored region shows the distribution of the adopted model library of complex star formation and metal enrichment histories (see Sec. \ref{['sec:stelpop_method_library']}), color-coded for r-band light-weighted age (left-most panels) and for r-band light-weighted stellar metallicity. Symbols are as in Fig.\ref{['fig:UVJ_SSFR']}: galaxies are distinguished into quiescent (magenta) and star-forming (blue) based on a mass-dependent threshold in specific SFR; empty symbols indicate galaxies belonging to the silver sample, i.e. with optimal combinations of indices, while filled symbols refer to the golden subsample (with $\rm S/N>20$); dots display all the galaxies in the LEGA-C sample with measurements of the two indices in each panel. In each panel only silver/ golden galaxies for which the corresponding couple of indices has been used in the fit are shown (therefore all the silver/ golden galaxies shown lie within $3\sigma$ of the model region). In each panel we report the number of galaxies shown among those in the full LEGA-C/ silver/ golden samples. The error bars display the median uncertainty on each index for the golden (thick black) and the silver galaxies (thin grey).
• Figure 4: Comparison between observed and best-fit predicted index strengths for the absorption features considered in the BaStA fit as age-sensitive diagnostics. Grey filled circles show the full primary LEGA-C catalog for which the index is measured; black circles highlight galaxies in the silver sample; magenta/blue circles highlight quiescent/star-forming galaxies in the golden sample. Filled symbols indicate galaxies for which the index has been used in the fit. The thin/thick errorbars indicate the median observational error for silver/ golden galaxies. The median offset, rms scatter and the median uncertainty are reported in each panel for the silver sample. The right-hand panels show the distribution of the difference between the absorption index strength predicted by the best-fit model and the observed one, normalized by the observational error (regardless of whether the index was used in the fit). Grey histograms are for the whole LEGA-C sample (normalized to unit area); black/magenta/blue histograms refer to the silver/ golden Q/ golden SF (normalized to the silver sample). A gaussian centered on zero and with unit standard deviation is shown for reference (dot-dashed curve).
• Figure 5: Same as Fig. \ref{['fig:indx_obs_bf_age']} but for the absorption features considered in the BaStA fit as metallicity-sensitive diagnostics.