Atomic hydrogen reservoirs in quiescent galaxies at z = 0.4

TL;DR

This paper investigates neutral atomic hydrogen in quiescent galaxies at $\\langle z \\rangle = 0.36$ through 21-cm stacking of MIGHTEE-HI data. Using color-selected and $D_n4000$-based spectroscopic quiescence criteria, it finds HI is substantial in dusty QGs but largely absent in non-dusty ones, with dusty QGs having $M_{HI} \approx 0.6\,M_{HI}^{MS}$ at fixed $M_{\star}$ (roughly 40% below star-forming galaxies of the same mass) and color-selected QGs HI-poor by a factor of about 3. HI content additionally correlates with dust content, declines with increasing $Dn4000$, and depends on environment and morphology (spirals retaining more HI than spheroids; lower-density regions hosting more HI). These results support slow quenching and ongoing gas/dust retention in a significant fraction of QGs at this epoch, and motivate future SKA-scale surveys to improve statistics and enable direct detections.

Abstract

Context. Based on Local Universe observations, quiescent galaxies (QGs) host lower to no HI compared to star-forming galaxies (SFGs), but no constraints have been derived so far at higher redshift (z>0.1). Understanding whether QGs can retain significant HI reservoirs at higher z is crucial to refine quenching and gas accretion models and to constrain overall star formation efficiency at different epochs. Aims. We aim to probe HI in candidate QGs at intermediate redshifts (z=0.36) and to understand whether there exists a class of QGs retaining consistent HI reservoirs and which parameters (dust content, stellar mass, Dn4000, morphology, environment) effectively capture HI-rich QGs. Methods. We perform 21-cm spectral line stacking on MIGHTEE-HI data at z=0.36, targeting two different samples of QGs, defined by means of a color-selection criterion and a spectroscopic criterion based on Dn4000, respectively. We also perform stacking on subsamples of the spectroscopically-selected quiescent sample to investigate the correlation between the HI content and other galaxy properties. Results. We find that QGs with an IR counterpart (i.e., dusty galaxies) are found to host a substantial HI content, on average just 40% lower than SFGs. In contrast, color-selected QGs still hold HI, but lower than SFGs by a factor 3. Among dusty objects, we find morphology to have a mild impact on the atomic gas content, with spirals hosting approximately 15-30% more HI than spheroids. Environmental effects are also present, with low-density regions hosting galaxies that are HI-richer than in high-density ones, by approximately 30% for spirals and 60% for spheroids. We suggest that, in general, HI content is driven by several factors, including quenching mechanisms and ISM enrichment processes. Also, quiescent galaxies - and especially dusty systems - seem to yield HI more consistently than in the Local Universe.

Figures (12)

• Figure 1: Two-dimensional distributions of galaxy properties from three investigated samples. Upper panel: Scatter plot of the spectroscopic sample from Donevski2023 in the $D_n$4000-$\log{M_{\star}}$ plane. Red triangles indicate galaxies with an IR counterpart, while gray circles mark galaxies with no IR detection. Histograms above and right of the scatter plot show the distribution of these parameters for the two infrared populations, using the same color coding. Lower panel: Distribution of galaxies from the spectroscopic sample of Donevski2023, plotted in comparison to the MS galaxies shown in the left panel. Red points represent galaxies with an IR counterpart, while gray points represent galaxies without an IR counterpart. The sample is compared with an MS model by Popesso2023 (black solid line), while the dashed black line indicates a 0.3 dex scatter around the MS.
• Figure 2: Comparison of average HI content between SFGs and QGs. The solid purple line represents the $M_{\rm HI}$–$M_{\star}$ scaling relation for SFGs from Bianchetti2025. Blue circles show HI masses from color-selected QGs in two stellar mass bins: a circle at the high-mass end (significant detection) and a triangle at the low-mass end (upper limit only). Red squares indicate HI masses from stacked spectra of dusty, spectroscopically selected QGs in two mass bins. A gray triangle marks the upper limit from stacking non-dusty, spectroscopically selected QGs.
• Figure 3: Average HI fraction of dusty QGs in $D_n$4000 bins. The average $f_{HI{}}=M_{\rm HI}/M_{\star}$ is shown in $D_n$4000 bins (1.5<$D_n$4000<1.61, 1.61<$D_n$4000<1.8 and $D_n$4000>1.8), plotted against their average value. The dashed line indicates the best-fit linear model in both panels.
• Figure 4: Logarithm of the ratio between the average HI mass of dusty QGs and the equivalent HI contained at the same stellar mass in MS galaxies according to Bianchetti2025, plotted against the median overdensity $\delta$ for each galaxy bin. A dashed horizontal line at y=0 indicates the HI content in SFGs. Purple filled squares mark the HI estimates from stacking dusty galaxies in lower-density ($\delta$<1) and in higher-density ($\delta$>1) bins, irrespective of morphology; shaded bands indicate uncertainties. Red and teal dotted lines indicate the HI estimate from stacking spirals and spheroids, respectively, irrespective of environment, spanning across the $\delta$ range.
• Figure 5: Logarithm of the ratio between average HI mass of dusty QGs and equivalent HI mass in MS galaxies at the same stellar mas according to Bianchetti2025, plotted against the median overdensity $\delta$ for each galaxy bin. A dashed horizontal line at y=0 indicates the HI content of SFGs. Filled purple squares mark the HI estimates from stacking dusty galaxies in lower-density ($\delta$<1) and in higher-density ($\delta$>1) bins. Empty red stars represent spiral and irregular galaxies divided into the two separate $\delta$ bins. Empty teal circles indicate spheroid galaxies in the same $\delta$ bins.