The Milky Way Bulge Extra-Tidal Star Survey: NGC 6569

TL;DR

This study presents the first wide-field, medium-resolution spectroscopic survey of the bulge globular cluster NGC 6569 aimed at identifying extra-tidal debris and quantifying ongoing mass loss. By combining Blanco DECam Bulge Survey photometry, Gaia DR3 astrometry, and AAOmega spectroscopy, the authors identify 7 Grade A and 12 Grade B high-confidence extra-tidal candidates whose chemo-dynamical properties match the bound cluster population, implying active tidal stripping with a present-day mass-loss rate of roughly 5–14% of the cluster mass per Gyr. The analysis employs time-dependent Jacobi radii from barred Milky Way potentials (AGAMA) and mock tidal streams (Gala) to interpret tail morphologies, leading to a consistent picture of breathing tidal radii $r_J(t)$ and nascent leading/lagging tails feeding bulge field stars. The work provides quantitative evidence that bulge GCs contribute to the inner Galaxy’s stellar halo and red clump structure, and establishes a framework for future N-body modeling and extended MWBest surveys. Overall, the paper advances our understanding of globular cluster dissolution in the bulge and the chemical-kinematic tagging of tidal debris.

Abstract

We present spectroscopic evidence for tidal debris associated with the bulge globular cluster NGC 6569, based on medium-resolution (R ~ 11,000) Anglo-Australian Telescope spectra of 303 stars. Targets were selected using Blanco DECam Bulge Survey (BDBS) photometry and Gaia DR3 astrometry, spanning 7-30 arcmin (~1-5 rt, where rt is the King-model tidal radius) from the cluster center. Orbit-based modeling predicts a strongly time-variable Jacobi radius, with rJ ~ 8-11 arcmin near pericenter and ~18-22 arcmin near apocenter, so stars just outside rt can be unbound and feeding leading and lagging tidal tails. We identify 40 stars with kinematics and abundances consistent with previous, or borderline, cluster membership. The seven highest-quality candidates (S/N > 30) have mean [Fe/H] = -0.83 +/- 0.14 and [alpha/Fe] = +0.38 +/- 0.06 dex, matching the bound population. Interpreting these stars as recently stripped debris implies a present-day mass-loss rate of 1.0-1.6 solar masses per Myr, or 5.6 +/- 1.3% of the current cluster mass per Gyr. These results indicate ongoing tidal stripping of NGC 6569 and quantify its contribution to the bulge field. This paper is part of the Milky Way Bulge Extra-Tidal Star Survey (MWBest) and is our first detailed debris study of a massive bulge globular cluster.

Figures (10)

• Figure 1: (a) The 18,180 objects within a radius of 30 of the cluster center, at $RA=273.\degree412000$, $Dec=-31.\degree826444$ with BDBS i-band measurements between 10.4 and 19.8 mag., overlaid on a variable extinction map simion17kader23 in the BDBS i-band. (b) Gaia DR3 sources were queried from gaiadr3.gaia_source around the SIMBAD center of NGC 6569 and filtered to have measured proper motions, $G<19$, and $\mathrm{RUWE}<1.4$. Three subsets are shown: (i) an inner $5\arcmin$ circle around the cluster center ("cluster"; blue points), (ii) a $15\arcmin$-$60\arcmin$ annulus ("bulge/field"; gray density histogram, randomly sub-sampled to $1.2\times10^{5}$ rows via random_index for speed), and (iii) a $2\arcmin$ core used only to estimate the cluster mean motion. The green star marks the median cluster proper motion from the $2\arcmin$ core. Tiny black dots show proper motions from our BDBS catalog overplotted on the same axes. Axes are $\mu_{\alpha*}\equiv \mathrm{d}\alpha/\mathrm{d}t \cos\delta$ versus $\mu_{\delta}$, in $\mathrm{mas\,yr^{-1}}$. No parallax or CMD cleaning is applied here, so the $5\arcmin$ (blue) selection includes some field contaminants.
• Figure 2: The BDBS differentially-dereddened CMD for $i_0\; vs.\; (u-i)_0$ is shown. The small, gray-filled circles are the stars within 12 of the center of NGC 6569. The following objects are identified within 5 tidal radii of the center of the cluster: stars within 0.95$^\circ$ of the cluster center; stars must have a PM within 1 mas$\,yr^{-1}$ in RA and within 1.5 mas$\,yr^{-1}$ in Dec of the main cluster PM: a box drawn between ($\sim -4.125 \pm 1.0$ mas$\,yr^{-1}$, $\sim -7.354 \pm 1.5$ mas$\,yr^{-1}$). Additional constraints were applied: objects with BDBS u-magnitudes with uncertainties $<0.04\; mag.$ were selected if they had detections in the BDBS g- and i-bands also. The red open squares were identified as targets for AAT/AAOmega spectroscopy. Blue triangles are flagged as likely foreground objects, and black-filled squares are identified as RRLs. The isochrone shown is from the MIST models choi16, $\textrm{[Fe/H]} = -0.75$ dex, scaled for $[ \alpha /Fe] = +0.3$ dex external to MIST, with an age of 13 Gyr.
• Figure 3: An example spectrum is shown with the regions used for the spectral analysis process identified. This is an extracted, processed, and normalized spectrum for AAT/AAOmega fiber #215. The normalized spectrum is shown in gray. The centroid of each CaT feature is shown as a red dotted line, which has been shifted back to the rest frame. The specutils routine was used to find lines by the derivative method, and the results were averaged for the three CaT lines. The blue line is the standard RGB spectrum which best fits the target star using the SP_Ace code boeche16boeche21, and the green line is the fitted region, not using the calcium triplet, and avoiding the sky-lines at longer wavelengths than 8800Å. The results for this star are: $\mathrm{[Fe/H]}_{\rm DP}$$=-0.92\pm 0.09$ dex, $\mathrm{[Fe/H]}_{\rm SP}$$=-0.83\pm 0.08$ dex, and $[ \alpha/Fe] =+0.27\pm 0.27$ dex.
• Figure 4: (a) A map of the RA vs. Dec positions of the 303 spectroscopic targets (2 fibers did not yield measurable spectra) in and around NGC 6569 (shaded circles), along with some known RR Lyrae stars (OGLE) with cataloged RVs (filled triangles). The circles are color-coded by RV (see color bar on the right). The 100 objects from J18 are shown as green crosses in the cluster center. The gray circles are successive tidal radii ortolani01valenti11 from the center of NGC 6569. We choose our search limit as $-25>RV>-65$ km s$^{-1}$ , yielding 58 objects (black-edged green circles with black pluses). (b) The relative number of stars per bin is shown in this histogram of the cluster RV-range yielded by J18, from $-30>RV>-63$ km s$^{-1}$ . The 100 stars from J18 were added to our 303 targets to find the cluster (green bars). The AAT sample only is shown as the black line (no shading). The Besançon Galactic population models for our FOV and PM limits are seen in mauve.
• Figure 5: Comparisons between the calibration methods of obtaining [Fe/H] from the CaT lines dias20 with the SP_Ace code results. (a) A histogram of the J18 sample of 98/100 stars is shown in green, which were matched with BDBS photometry. For these 98 stars, using the luminosity correction in the i-band, we find: $\textrm{[Fe/H]}_{CaT}= -0.84\pm 0.17$ dex. (b) The 58 stars in the RV-selected sample ($-25>RV>-65$ km s$^{-1}$ ) are shown as the black-line histogram (no fill), calibrated using dias20 method. For all objects: $\textrm{[Fe/H]}_{DP} = -0.91\pm 0.40$ dex. Targets with $3500<T(K) <7500$ were fit with the SP_Ace code: $\textrm{[Fe/H]}_{SP} = -0.78\pm 0.20$ dex. The dotted line is the Gaussian fit to the P23 data, presented in crociati23: $\textrm{[Fe/H]}_{CaT} = -0.90\pm 0.20$ dex. (c) Comparing the two fits for the overlap between the samples (all stars are faint green and orange filled circles). The red line is the 1:1 fit of the CaT-method and the SP_Ace code, which fits the stars with both $\mathrm{[Fe/H]}_{\rm DP}$ and $\mathrm{[Fe/H]}_{\rm SP}$ are limited to values between -0.5 and -1.1 dex (40 objects; brighter green and orange points).