Diffuse Dwarf Galaxies in Galaxy Clusters: I. Stellar Populations and Radial Gradients

Abstract: We use Keck/KCWI spectroscopy to study one ultra-diffuse galaxy (UDG) and five Nearly-UDGs (NUDGEs) in the Perseus cluster, together with an additional UDG in the Coma cluster. As the first paper in a series, we focus on the global and radial stellar population properties of our sample. We find that these galaxies host intermediate-to old stellar populations, with typical ages of ~7 Gyr, low metallicities ([M/H]$\simeq$ -0.9 dex), and enhanced [Mg/Fe] abundances (~0.3 dex), consistent with previous studies. Six galaxies lie within the scatter of the present-day mass-metallicity relation (MZR), whereas the Coma UDG (DF11) is more consistent with the MZR of high-z galaxies (z ~ 2). We find no strong correlation between global stellar population properties and cluster infall parameters, suggesting that any environmental impact is not easily traceable through integrated stellar populations. We go one step further and measure radial gradients for three galaxies. Two show flat age and mildly negative metallicity gradients, similar to classical dwarfs, while one shows a rising metallicity profile as recently found in other UDGs. Comparing with classical dwarfs, we find a continuous correlation between metallicity gradient and globular cluster (GC) richness, where more GC-rich systems tend to show rising profiles. We propose that preferential tidal disruption of GCs in the inner regions of galaxies naturally produces rising metallicity profiles, unlike GC-poor classical dwarfs. This mechanism, potentially coupled with strong stellar feedback from early concentrated star formation, may explain the unusual rising metallicity profiles observed in GC-rich UDGs/NUDGEs.

• The paper presents Keck/KCWI integral-field spectroscopy of seven diffuse dwarf galaxies in clusters, revealing a range of star formation histories and global stellar properties.
• High-SN galaxies show unique radial metallicity gradients, which correlate with GC richness;'DF11' shows a positive gradient, while classical dwarf galaxy 'H332' and 'W1' follow outside-in formation models.
• The authors identify a positive correlation between metallicity gradients and GC richness, proposing mechanisms involving tidal disruption of metal-poor GCs and early feedback-driven supershells.

Sample and observations

This paper presents Keck/KCWI integral-field spectroscopy of seven diffuse dwarf galaxies in dense cluster environments: six in the Perseus cluster (H332, H410, H586, H629, H662, W1) and one in Coma (DF11). Two objects satisfy the "bona fide" UDG criteria of van Dokkum et al. ($\mu_{0,g} \ge 24$ mag arcsec$^{-2}$, $R_{\mathrm{e}} \ge 1.5$ kpc); the remaining five are NUDGEs. A key design feature is GC-richness diversity: five galaxies are GC-poor and two (H410, H662) are GC-rich with $N_{\rm GC} \gtrsim 20$, whereas most prior spectroscopic UDG gradient samples are dominated by GC-rich systems. Observations used the KCWI Medium slicer with BL or BH3 gratings on the blue side ($\sigma_{\rm inst}$ = 67 or 13 km s$^{-1}$) plus RH2 on the red arm; only cloudless blue-side frames were retained.

Global stellar populations

Full-spectrum fitting with pPXF against E-MILES/BASTI templates yields a sample mean mass-weighted age of $7.1 \pm 2.0$ Gyr (range ~3.5–10 Gyr), mean [M/H] $= -0.9 \pm 0.2$ dex, and mean [Mg/Fe] $\simeq 0.3$ dex from two independent line-index methods that agree within uncertainties. Two NUDGEs show extended star formation histories with quenching within the last ~3 Gyr; the rest quenched earlier, with t$_{90}$ exceeding ~7 Gyr for several systems. The authors explicitly caution that low-S/N, short-baseline BH3 spectra may bias ages toward intermediate values — an important caveat for the three young NUDGE measurements.

On the mass–metallicity plane, six galaxies fall within the scatter of the present-day dwarf MZR, but DF11 ([M/H] $^{-2}$0 dex, [Mg/Fe] $^{-2}$1 dex) lies closer to the theoretical $^{-2}$2 MZR (deviating by only ~0.05 dex), reinforcing its candidacy as a failed galaxy despite its nominally GC-poor classification ($^{-2}$3, consistent with GC-rich within errors).

Environmental dependence

Using the infall parameter $^{-2}$4 following Tang et al., no correlation is found between age or metallicity and infall time. Only [Mg/Fe] shows a weak negative relation with $^{-2}$5 (Spearman coefficient $^{-2}$6, $^{-2}$7), consistent with $^{-2}$8-abundance acting as a chemical clock for earlier-infall systems. The absence of age/metallicity trends is attributed partly to pre-infall quenching for some systems and to the statistical limitations of projected phase-space diagnostics, including backsplash contamination.

Radial gradients

For the three highest-S/N targets (H332, W1, DF11), annular pPXF fits yield:

Galaxy	$^{-2}$9	$R_{\mathrm{e}} \ge 1.5$0
W1	$R_{\mathrm{e}} \ge 1.5$1	$R_{\mathrm{e}} \ge 1.5$2
H332	$R_{\mathrm{e}} \ge 1.5$3	$R_{\mathrm{e}} \ge 1.5$4
DF11	$R_{\mathrm{e}} \ge 1.5$5	$R_{\mathrm{e}} \ge 1.5$6

W1 and H332 follow classical outside-in dwarf formation, while DF11 shows a significantly positive metallicity gradient, joining DF44, DFX1, and PUDG-R27 as UDGs deviating from classical dwarfs and TNG50 simulated UDGs. DF11 also provides only the second UDG [Mg/Fe] gradient measurement after DF44, found to be flat.

The GC-richness connection

The central result is a statistically significant positive correlation between metallicity gradient and $R_{\mathrm{e}} \ge 1.5$7 (Spearman coefficient 0.55, $R_{\mathrm{e}} \ge 1.5$8; slope $R_{\mathrm{e}} \ge 1.5$9), with no corresponding trend for age gradients. Systems with high $N_{\rm GC} \gtrsim 20$0 preferentially exhibit rising metallicity profiles and lower global metallicities. Tidal stripping is rejected as the dominant explanation because none of the sample shows tidal disturbance signatures and metallicities remain at or below the MZR.

The paper advances two complementary mechanisms: (i) preferential tidal disruption of metal-poor GCs in inner galactic regions deposits metal-poor stars centrally, inverting gradients — a toy model requiring ~20% stellar mass contribution from disrupted GCs to raise inner metallicity by 0.1 dex, comparable to the ≥25% disrupted-GC contribution estimated for the Milky Way inner halo; and (ii) Mori-style feedback-driven outward-propagating supershells from an early concentrated burst that also formed the GCs, building stars in progressively enriched gas. Both plausibly combine, with the strongest effects in the most GC-rich UDGs.

Limitations

The gradient sample comprises only three galaxies, and the environmental analysis is limited by projected phase-space degeneracies and small-number statistics; the paper itself notes that firmer infall correlations require much larger UDG samples and that simulated gradient comparisons carry subgrid feedback uncertainties. Whether GC-rich dEs share the same gradient mechanism remains untested, since most literature dEs are GC-poor. Extreme systems such as NGC5846_UDG1 and DGSAT I have not yet been spatially resolved spectroscopically.

Conclusion

By adding five GC-poor and two GC-rich cluster UDGs/NUDGEs to the spectroscopic sample, this study confirms that most cluster diffuse dwarfs resemble classical dwarfs in their global populations while establishing a continuous, significant link between GC richness and rising metallicity gradients. It supports a composite-formation picture of cluster UDGs in which GC disruption and early feedback jointly shape resolved chemical structure, leaving open whether detailed GC-disruption simulations can reproduce the full observed range of positive gradients.