$S^5$: Tidal Disruption in Crater 2 and Formation of Diffuse Dwarf Galaxies in the Local Group

TL;DR

This study uses the S^5 spectroscopic survey to dissect Cra2 and its tidal stream, combining a Bayesian mixture model with tailored N-body simulations in MW+LMC potentials to probe Cra2’s dark matter halo. The analysis reveals a dynamically cold central system (σ_v,gal ≈ 2.5 km s^-1) coexisting with a hotter stellar stream (σ_v,str ≈ 5.7 km s^-1) and a substantial on-sky velocity gradient along the stream, implying strong tidal disruption. The measured stream-to-remnant velocity-dispersion ratio (≈2.3) disfavors a fiducial cuspy NFW halo and favors either a low-concentration cusp or a cored halo with a small core radius, while Cra2 remains metal-poor (⟨[Fe/H]⟩ ≈ -2.16) and below the Local Group MZR after accounting for mass loss. These results, in tandem with analogous diffuse dwarfs Antlia 2 and Andromeda 19, suggest tidal processes and alternative DM density profiles play a key role in the formation of ultra-diffuse Local Group dwarfs, challenging conventional CDM expectations and motivating further simulations and chemical-abundance studies.

Abstract

We present results of a spectroscopic campaign around the diffuse dwarf galaxy Crater 2 (Cra2) and its tidal tails as part of the Southern Stellar Stream Spectroscopic Survey ($S^5$). Cra2 is a Milky Way dwarf spheroidal satellite with extremely cold kinematics, but a huge size similar to the Small Magellanic Cloud, which may be difficult to explain within collisionless cold dark matter. We identify 143 Cra2 members, of which 114 belong to the galaxy's main body and 29 are deemed part of its stellar stream. We confirm that Cra2 is dynamically cold (central velocity dispersion $2.51^{+0.33}_{-0.30}\,{\rm km\,s^{-1}}$) and also discover a $\approx$7$σ$ velocity gradient consistent with its tidal debris track. We separately estimate the stream velocity dispersion to be $5.74^{+0.98}_{-0.83}\,{\rm km\,s^{-1}}$. We develop a suite of $N$-body simulations with both cuspy and cored density profiles on a realistic Cra2 orbit to compare with $S^5$ observations. We find that the velocity dispersion ratio between Cra2 stream and galaxy ($2.30^{+0.41}_{-0.35}$) is difficult to reconcile with a cuspy halo with fiducial concentration and an initial mass predicted by standard stellar mass$-$halo mass relationships. Instead, either a cored halo with relatively small core radius or a low-concentration cuspy model can reproduce this ratio. Despite tidal mass loss, Cra2 is metal-poor ($\langle \rm[Fe/H]\rangle=-2.16\pm0.04$) compared to the stellar mass$-$metallicity relation for its luminosity. Other diffuse dwarf galaxies similar to Cra2 in the Local Group (Antlia 2 and Andromeda 19) also challenge galaxy formation models. Finally, we discuss possible formation scenarios for Cra2, including ram-pressure stripping of a gas-rich progenitor combined with tides.

Figures (7)

• Figure 1: Left: absolute $V$-band magnitude ($M_V$) vs. size diagram for Local Group dwarf galaxies pace2025lvdb. The vertical axis corresponds to the half-light major axis $R_h = {R_{1/2}}\xspace(1-\epsilon)^{-1/2}$, where ${R_{1/2}}\xspace$ is the circularized half-light radius and $\epsilon$ is the ellipticity. The corresponding stellar masses ($M_\star$) are computed as in Equation \ref{['eq:mstar']}. Diagonal lines exhibit constant values of surface brightness following annotations. Diffuse dwarfs Crater 2 (green '$\star$' symbol), Antlia 2 (yellow 'X'), and Andromeda 19 (purple '$+$') are highlighted. Milky Way satellites, M31 satellites, and Local Group field dwarf galaxies are plotted as blue squares, orange triangles, and gray circles, respectively. Right: $V$-band surface brightness ($\mu_V$) vs. size ($R_h$). Ultra-diffuse galaxies (UDGs) in the Coma cluster vanDokkum2015udgs are shown as the red dots. White circles with red edges represent other UDGs from a spectroscopic compilation gannon2024udgs. Gray dots are typical galaxies from the Sloan Digital Sky Survey data release 7 within the redshift range $0.01 \leq z \leq 0.03$simard2011sdss.
• Figure 2: On-sky stellar spatial distribution around the Crater 2 (Cra2) galaxy. Green circles with black edges are $S^5$ Cra2 members with a probability $P> 50\%$ from our mixture model (Section \ref{['subsec:gmm']}), including remnant body and stream. Other Cra2 candidates ($20\% < P \leq 50\%$) are shown as the smaller green dots. Variable RR Lyrae stars in the Cra2 stream are displayed as the dark blue circles with white edges coppi2024cra2. The directions of both "far side" ($\alpha \lesssim 177.5^\circ$) and "near side" ($\alpha \gtrsim 177.5^\circ$) of the stream are annotated. The 2dF$+$AAOmega fields observed as part of $S^5$ are presented as the dashed black contour. The thin dotted lines mark 3, 5, 7, 10, and 15 ${R_{1/2}}\xspace$ around Cra2, where ${R_{1/2}}\xspace = 31.2$ arcmin with a null ellipticity Torrealba2016cra2. The Cra2 core-base$N$-body model star particles (Section \ref{['subsec:nbody']}) are the gray dots in the background. Finally, the red arrow and dashed line exhibit the Cra2 proper motion vector in Galactic standard of rest frame and its past orbit used for the $N$-body simulations, respectively.
• Figure 3: Heliocentric distances as a function of stream longitude $\phi_1$ for the Crater 2 (Cra2) system. Dark blue symbols with white edges are RR Lyrae stars associated with the Cra2 stream coppi2024cra2. The simulated Cra2 star particles from the core-base model (Section \ref{['subsec:nbody']}) are shown as the small gray dots. The gray and black lines represent the $-4.9\,{\rm kpc}\,{\rm deg}^{-1}$ and $-5.4\,{\rm kpc}\,{\rm deg}^{-1}$ heliocentric distance gradients predicted by the core-base and cusp-base simulations, respectively.
• Figure 4: Mixture modeling fit results. Top panel: Crater 2 (Cra2) stream coordinates $(\phi_1, \phi_2)$ distribution (Section \ref{['subsec:coords']}). The far (trailing arm) and near (leading) sides of the stream are annotated. Middle: linear proper motion gradients as a function of stream longitude $\phi_1$. Bottom left: linear Galactic standard of rest line-of-sight velocity ($v_{\rm gsr}$) gradient. Bottom right: radial metallicity ($\rm [Fe/H]$) gradient. In all panels, Cra2 candidates ($P>20\%$) in the $S^5$ DR2 footprint are color-coded by their membership probabilities. Non-members that have also been observed by $S^5$ are plotted as gray dots. Red dashed lines represent the mean values from our best-fit mixture model while the thin green lines show 200 realizations of our fit sampled from our posterior chains. The black solid lines are linear fits to our core-base simulated stream track (Section \ref{['subsec:nbody']}). Red giant-branch Cra2 stream candidates from coppi2024cra2 are plotted as black open squares. Those with available $S^5$ DR2 data, but rejected as true members, are displayed as black 'X' symbols. In all panels, blue dashed lines delineate the formal boundary between galaxy and stream in our mixture modeling analysis (Section \ref{['subsec:gmm']}).
• Figure 5: Dereddened DECam $(g-r)_0$ vs. $M_{r,0}$ color--magnitude diagram covering the red giant branch (RGB) with photometry from DELVE DR2 DELVEdr2 for stars in the $S^5$ DR2 footprint with Crater 2 (Cra2) membership probability $P>50\%$ (green circles with black edges) and $20\%<P\leq50\%$ (green dots). The absolute $r$-band magnitude is computed assuming the expected distance gradient from our Cra2 core-base model (Section \ref{['subsec:nbody']}). The coppi2024cra2 Cra2 RGB candidates are plotted as the open squares, with the rejected ones having an 'X' symbol on top. Red lines are (12.5 Gyr, $[\alpha/{\rm Fe}] = +0.4$) Dartmouth isochrones dotter2008 with ${\rm [Fe/H]}\xspace = -2.5$ (solid), $-2.0$ (dashed), and $-1.5$ (dotted).