A$^3$COSMOS: The dust content of massive quiescent galaxies and its evolution with cosmic time

TL;DR

This study probes the dust content and cold ISM evolution of massive quiescent galaxies from $z=0.5$ to $3$ by stacking ALMA data in the uv plane for a mass-complete sample ($\log(M_*/M_{\odot})\ge 10.8$). Using CIGALE to derive star-formation histories and time-since-quenching, the authors find that most quenching occurs around $z_q\sim1.3$, with a large fraction of high-redshift QGs recently quenched. They derive stringent upper limits on the dust-to-stellar-mass fraction, $\langle f_{\mathrm{dust}}\rangle_{\mathrm{pop}}\lesssim 4\times10^{-4}$, implying a much lower dust (and gas) content than star-forming galaxies and supporting rapid quenching via gas removal or consumption within $\sim$600 Myr. The results indicate that most gas and dust are expended or expelled early in quiescence, with little re-accretion, and highlight the effectiveness of high-resolution ALMA stacking in constraining ISM evolution in the quenched galaxy population. Future work with larger samples and deeper data will sharpen the inferred dust removal timescales and the role of re-accretion in quenching.

Abstract

We study the dust content of massive ($\log(M_*/M_{\odot})\geq10.8$) quiescent galaxies (QGs) at redshifts $z=0.5-3$ to place constraints on the evolution of their cold interstellar medium (ISM) and thereby obtain insights into the processes of galaxy quenching throughout cosmic time. We used a robust sample of 458 colour-selected QGs covered by the A$^3$COSMOS+A$^3$GOODSS database to perform a stacking analysis in the $uv$ domain and measured their mean dust masses from their stacked sub-millimetre luminosities. We used the CIGALE spectral energy distribution fitting code to obtain star formation histories and infer the time since quenching for all the QGs in our sample. We used this information to gain insight into the time evolution of the dust content after quenching. Most QGs in our sample quenched around a redshift of $z\sim1.3$, following the peak of cosmic star formation. The majority of QGs observed at $z>1$ are recently quenched (i.e. quenched for no longer than 500 Myr), whereas the majority of QGs observed at $z<1$ have already been quenched for a significant amount of time ($\gtrsim1$ Gyr). This implies that high-redshift galaxies ($z\gtrsim2$) are ideal for studying the mechanisms of quenching and its effects on the ISM, while lower-redshift galaxies are more suitable for studying the long-term effects of the QG environment on their ISM. We obtain upper limits on the dust mass fraction of the QG population that indicate a lower dust content than what was found by earlier stacking studies, and significantly lower (by a factor of $\sim2-6$) than that of normal star-forming galaxies. We also place constraints on the initial gas fraction right after quenching. We find that within the first $\sim600$ Myr after quenching, QGs already lose on average $\gtrsim70\%$ of their cold ISM. Our findings support a gas consumption or removal scenario acting on short timescales.

Figures (10)

• Figure 1: Flow diagram of the selection of a robust QG sample. Blue panels denote general selections, orange panels the selection of QGs, red panels further cleaning steps, and green panels the number of QGs after every step.
• Figure 2: Normalised distribution of CIGALE-inferred Bayesian parameters of the parent sample of 4583 QGs and the 460 QGs in our A$^3$COSMOS/A$^3$GOODSS sample: SFR fraction ($r_{\mathrm{SFR}}$), stellar mass ($M_*$), time since quenching ($t_{\mathrm{q}}$), and ($\Delta$MS) before quenching and at the time of observation. We note that, due to the re-estimation of $M_*$ with CIGALE, a small number of galaxies fall slightly below the initial selection threshold of $\log(M_*/M_{\odot})=10.8$.
• Figure 3: Characterisation of our sample of A$^3$COSMOS/A$^3$GOODSS QGs and the parent sample. Top panel: Time since quenching as a function of the observed redshift. The median error is shown on the right. Solid black lines indicate the locus of galaxies that quenched at the same time, from $z_{\mathrm{q}}=0.5$ to 5 in steps of 0.5. The orange density map shows the distribution of the parent sample galaxies. The grey shaded areas in the background and dashed grey lines indicate the bins of quenching redshift and time since quenching used in Sect. \ref{['subsec_dustdestruction']}. Middle panel: Distribution of quenching redshift ($z_{\mathrm{q}}$). Bottom panel: Fraction of recently quenched galaxies ($t_{\mathrm{q}}\leq500$ Myr). The points and horizontal error bars denote the median redshift and the 16th and 84th percentiles of all galaxies in each bin, respectively. The black line shows the fraction of recently quenched massive early-type galaxies as a function of redshift from gobat20. The dotted blue line and shaded area show our toy model based on the stellar mass functions of davidzon17.
• Figure 4: Results of our stacking of four different redshift bins: $z_{\mathrm{obs}}=0.5-1$, $1-1.5$, $1.5-2$, and $2-3$. Top panel: Cutouts of the map centre. We show the profile of the synthesised beam as a dashed white shape in the bottom left, and the number of individual galaxies stacked in the bottom right. The dashed red circle in the centre shows the aperture used to measure the source emission. The colour scaling is set to reflect the signal-to-noise ratio of each pixel. Bottom panel: Dust-to-stellar mass ratio ($f_{\mathrm{dust}}$) of our stacks (green diamonds) at the weighted mean redshift of each stack in comparison to the $f_{\mathrm{dust}}$ measurements of massive QGs from the literature; grey symbols show individual measurements from dust (grey hexagons) and CO (grey squares), yellow symbols the results from stacking studies, and the dashed yellow and dash-dotted black lines the predictions from data fitting. The empty diamonds show our upper limits when assuming a lower dust temperature of 17 K. The dashed blue lines and shaded regions show the $f_{\mathrm{dust}}$ evolution of massive MS galaxies from a3cosmos_2 and wang22.
• Figure 5: Dust mass at the time of observation divided by the dust mass of MS galaxies at the time of quenching as a function of time since quenching for three different bins of quenching redshift (green filled symbols). The empty symbols show the upper limits when assuming a lower dust temperature of 17 K. The yellow, orange-hatched, and red-chequered regions show the dust removal timescales from michalowski19, lesniewska23, and nadolny24, for three different fractional initial gas fractions ($f_{0,\mathrm{gas}}$). The empty purple symbol marks the $f_{0,\mathrm{gas}}$ from the fit of gobat20. Light blue and dark blue symbols show dust mass evolution with time since quenching from donevski23 for spiral and elliptical QGs at $z<0.7$, respectively, where we assume a 30% error on the dust mass, and the horizontal error bars denote the bin size of donevski23.