VVV CL001: Likely the Most Metal-poor Surviving Globular Cluster in the Inner Galaxy

The inner ¹⁵ Milky Way (MW) contains a significant population of globular clusters (GCs) with a wide range of metallicities (−2.37 < [Fe/H] ≲ 0; Harris 1996, 2010 edition), most of which are still poorly explored due to the large foreground extinction and high field-star densities that complicate the analysis, especially at low Galactic latitudes. These limitations have been mitigated with wide field near-infrared imaging such as the Vista Variables in the Via Lactea survey (VVV), and its extension the VVVX survey (Minniti et al. 2010; Smith et al. 2018), which has expanded the family of Galactic GCs to more than 300 candidates by the inclusion of objects in the inner Galaxy (see e.g., Minniti et al. 2017a, 2017b; Palma et al. 2019; Garro et al. 2020), including those in the bulge (D. Geisler et al. 2021, in preparation).

In this context, the high-resolution (R > 22,0000) capabilities of the near-IR multifiber spectrographs of the Apache Point Observatory Galactic Evolution Experiment (APOGEE-2; Majewski et al. 2017) allow measurement of new parameters (radial velocity, metallicity, and detailed chemical abundances for many species) with high precision for a large number of Galactic GCs in a homogeneous way (see, e.g., Mészáros et al. 2015, 2020; Schiavon et al. 2017; Fernández-Trincado et al. 2019d, 2020e; Masseron et al. 2019).

Very metal-poor GCs are preferentially associated with the oldest components of galaxies, and are often used as cosmic clocks to track the enrichment history of their host galaxy (Geisler et al. 1995). Thus, beyond the intrinsic scientific value of identifying such ancient objects, the measurements of their chemical composition can therefore provide insight into the build-up of the chemical elements in the earliest epoch of the Milky Way.

To date, the most metal-poor MW GC known is ESO280-SC06 (located ∼15 kpc from the Galactic center), with an estimated metallicity of [Fe/H] = −${2.48}_{-0.11}^{+0.06}$ (Simpson 2018), and associated to the Gaia–Enceladus–Sausage (GES; Massari et al. 2019). Here, we report another possible extreme case, VVV CL001, which becomes possibly the most metal-poor GC known inside the Sun's Galactocentric distance, and likely the most metal-poor GC in the Galaxy. It has survived the strong tidal field of the inner MW during its interaction, and its existence implies that GCs below that of the empirical metallicity floor (e.g., Larsen et al. 2020; Wan et al. 2020) have been formed, but only a few exceptional cases have survived during Galactic evolution.

In this Letter, we make use of APOGEE-2 spectra to provide the first spectroscopic study of VVV CL001.

We use an interim release data product of the APOGEE-2 survey (Majewski et al. 2017) part of the Sloan Digital Sky Survey IV (SDSS-IV; Blanton et al. 2017) that includes data taken after the release of DR16 (Ahumada et al. 2020), to investigate, for the first time, the chemical composition of VVV CL001.

The (300-fiber) APOGEE instruments are high-resolution (R ∼ 22,500), near-infrared spectrographs (Wilson et al. 2019) observing all the components of the MW (halo, disk, and bulge/bar) from the Northern Hemisphere on the 2.5 m telescope at Apache Point Observatory (APO, APOGEE-2N; Gunn et al. 2006) and from the Southern Hemisphere on the Irénée du Pont 2.5 m telescope (Bowen & Vaughan 1973) at Las Campanas Observatory (APOGEE-2S). Each instrument records most of the H band (1.51–1.69 μm). See Nidever et al. (2015), García Pérez et al. (2016), and Holtzman et al. (2018) for details regarding the data reduction process and standard stellar parameter estimates. As of 2020 February, the dual APOGEE-2 instruments have observed some ∼600,000 sources across the MW, targeting these stars according to methods described in Zasowski et al. (2013, 2017), with updates to the targeting plan described in R. Beaton et al. (2021, in preparation) and F. Santana et al. (2021, in preparation).

The present study focuses on VVV CL001, a GC discovered by the VVV survey (see, e.g., Minniti et al. 2011). This GC lies in the direction of the Galactic bulge, and is strongly dominated by large foreground extinction, E(B − V) ≳ 2.0 mag (Minniti et al. 2011), hampering the observations of this object in the optical bands.

VVV CL001 stars fall on the same APOGEE-2 plug-plates as those associated with the very nearby GC UKS 1, as shown in panel (a) of Figure 1 (see also Fernández-Trincado et al. 2020e, for details regarding to UKS 1); however, the high radial velocities (RVs) of potential VVV CL001 members allow us to cleanly distinguish the UKS 1 sources from VVV CL001 stars. Thus, two high-confidence VVV CL001 members, based on proper motions (see panel (b) of Figure 1) from Gaia Early Data Release 3 (Gaia EDR3: Gaia Collaboration et al. 2020), location in the color–magnitude diagrams (CMDs), and RVs from APOGEE-2 were identified ≲01 from the cluster center.

Zoom In Zoom Out Reset image size

Panels (c) and (d) of Figure 1 show that both our derived [Fe/H] and RV of the VVV CL001 stars are clearly distinct as compared to the foreground and background stars (hereafter, field stars). The [Fe/H] and RVs are at least 1.5 dex and ∼150 km s⁻¹ offset from the field stars, respectively. Both of these offsets are very large and imply that VVV CL001 is a truly extreme GC. The two potential cluster members from APOGEE-2 are red giant branch (RGB) stars close to the tip of the giant branch as shown in panel (e) of Figure 1.

In order to have a self-consistent method for age derivation via statistical isochrone fitting, we use the SIRIUS code (Souza et al. 2020), and the most probable cluster members in the VVV catalog located inside 15 from the cluster center and that have proper motions compatible with that of VVV CL001, as well as those sources with RVs information (see Section 4), which have been marked as open triangle (for Baumgardt's data set) and square (for APOGEE-2 data set) symbols in panel (e) of Figure 1. Due to the quality of our data, we applied some assumptions to obtain an age distribution: a uniform prior in age between 1 and 15 Gyr combined with a slow drop above the age of the universe (13.7 Gyr; Planck Collaboration et al. 2016); the metallicity was varied around the value determined with high-resolution spectroscopy in the present work; the isochrone is limited to $\mathrm{log}$ g < 4.5 representing the RGB region. We dereddened and extinction-corrected the VVV+2MASS Ks_s and J − K_s colors with the bulge-specific reddening maps from Gonzalez et al. (2011, 2012) assuming the reddening law of Cardelli et al. (1989). We also noticed that in an ∼2' × 2' area the differential reddening across the field does not affect the CMD of VVV CL001, with a negligible variation of 0.03 mag in K_s . Finally, we adopted the Dartmouth Stellar Evolutionary Database (DSED; Dotter et al. 2008) isochrones with [α/Fe] = +0.4 and canonical helium. Panels (e) and (f) of Figure 1 present the best isochrone fits in the K_s versus (J − K_s ) CMD. Our fit provides a reasonable solution both in the overplotted isochrone (panel (e)) and the posterior distributions of the corner plot (panel (f)). To represent the distributions, we adopt the median as the most probable value and the uncertainties calculated from the 16th and 84th percentiles. We found an age of ${11.9}_{-4.05}^{+3.12}$ Gyr and a probable distance of $\sim {8.22}_{-1.93}^{+1.84}$ kpc. Also, we want to stress that without the adopted assumptions, the internal error of the age determination could increase and not give a clear age distribution. Consequently, our probable solutions, within 1 − σ, fit well the central part of the CMD, providing confidence that the age estimate is a reasonable determination for VVV CL001.

As the ASPCAP/APOGEE-2 pipeline (García Pérez et al. 2016) does not provide [Fe/H] and [X/Fe] determinations for VVV CL001 stars, we followed the same technique as described in Fernández-Trincado et al. (2019a, 2019b, 2019c, 2019d, 2020a, 2020b, 2020c, 2020e), and carried out a consistent chemical-abundance analysis for the two VVV CL001 stars with the BACCHUS code (Masseron et al. 2016). The spectra of our sample, in general, have a signal-to-noise ratio (S/N) that is appropriate for elemental-abundance determinations; see Table 1. The atmospheric parameters (T_eff, $\mathrm{log}g$, and ξ_t ), metallicity ([Fe/H]), and elemental-abundance ratios ([X/Fe]) were derived with the BACCHUS code.

Table 1. Atmospheric Parameters and Elemental Abundances of Stars in VVV CL001

APOGEE-ID	2M17544233	2M17544268
	−2400536	−2400573
Teff (K)	4100	4200
$\mathrm{log}$ g (cgs)	0.50	0.50
ξt (km s−1)	1.81	1.69
S/N (pixel−1)	229	193
RV (km s−1)	−323.57	−326.16
RV-scatter (km s−1)	2.55	2.13
[N/Fe]	+1.23 ± 0.31	+0.23 ± 0.40
[O/Fe]	+0.41 ± 0.18	+0.39 ± 0.16
[Mg/Fe]	+0.04 ± 0.02	+0.15 ± 0.09
[Al/Fe]	−0.63 ± 0.11	−0.19 ± 0.09
[Si/Fe]	+0.26 ± 0.14	+0.33 ± 0.10
[Fe/H]	−2.47 ± 0.24	−2.44 ± 0.24

Note. The errors were determined in the same manner as described in Fernández-Trincado et al. (2020a) by varying the atmospheric parameters one at a time by the typical, albeit conservative, values of ΔT_eff = ±100 K, ${\rm{\Delta }}\mathrm{log}$ g = ±100 cgs, and Δξ_t = ±0.05 km s⁻¹. Thus, the reported uncertainties are defined as ${\sigma }_{\mathrm{total}}^{2}={\sigma }_{[{\rm{X}}/{\rm{H}}],{T}_{\mathrm{eff}}}^{2}+{\sigma }_{[{\rm{X}}/{\rm{H}}],\mathrm{log}g}^{2}+{\sigma }_{[{\rm{X}}/{\rm{H}}],{\xi }_{t}}^{2}+{\sigma }_{\mathrm{mean}}^{2}$.

Download table as:  ASCIITypeset image

Figure 2 shows the good quality of the APOGEE-2 spectrum compared to the best-fit synthetic spectra for selected atomic and molecule lines of the two stars in VVV CL001, from which the chemical-abundance ratios were determined. For each spectrum we were able to identify reliable lines for six chemical species (N, O, Mg, Al, Si, and Fe). The same figure also reveals the scarcely detectable Fe i lines. The resulting elemental-abundance ratios are listed in Table 1.

Zoom In Zoom Out Reset image size

The resulting chemical abundances are also displayed in Figure 3, and compared to four GCs at similar metallicity taken from the APOGEE-2 GC sample of Mészáros et al. (2020). We find that the two program stars exhibit an iron abundance ratio of −2.47 to −2.44, suggesting that VVV CL001 has a mean metallicity [Fe/H] = −2.45, with an uncertainty due to systematics of 0.24 dex, which makes this cluster possibly the most metal-poor GC identified so far within the Sun's Galactocentric distance, and with an extreme metallicity close to the apparent "floor" in the empirical metallicity distribution function for GCs in the MW and Local Universe (e.g., Geisler et al. 1995; Simpson 2018; Kruijssen 2019; Larsen et al. 2020; Wan et al. 2020). Note that VVV CL001 has the lowest Fe abundance among this sample, which includes the lowest metallicity GCs observed by APOGEE. With our limited sample, we do not find evidence of an intrinsic Fe-abundance spread.

Zoom In Zoom Out Reset image size

Regarding the α-elements (O, Mg, and Si), VVV CL001 displays a modest α-element enhancement, a clear signature of the fast enrichment provided by SNe II events, and compatible with other Galactic metal-poor GCs at similar metallicity (see Figure 3). Furthermore, no Mg–Al anticorrelation is evident in our small sample; the two VVV CL001 stars exhibit an aluminum deficit, which places them within the definition of first-generation stars according to the criteria developed by Mészáros et al. (2020). However, one star in our sample displays a very high enrichment in nitrogen, and is considered to likely belong to some of the families of second-generation ¹⁶ stars with low-aluminum enrichment (see, e.g., Mészáros et al. 2020), indicating the possible evidence for multiple stellar populations (MPs) in VVV CL001. However, we caution the reader that the error bars in either N or Al abundance (or both) are not properly understood or inconclusive; therefore, more cluster members need to be followed-up of in order to confirm or discard the existence of MPs in this VVV CL001.

We made use of the state-of-the-art MW model–GravPot16 to predict the orbital path of VVV CL001 in a steady-state gravitational Galactic model that includes a "boxy/peanut" bar structure.

For the orbit computations, we adopt the same Galactic model configuration, solar position, and velocity vector as described in Fernández-Trincado et al. (2020d), except for the angular velocity of the bar, for which we employed the recommended value of 41 km s⁻¹ kpc⁻¹ (Sanders et al. 2019).

The most likely orbital parameters and their uncertainties are estimated using a simple Monte Carlo approach. An ensemble of 10,000 orbits were calculated under variations of the observational parameters according to their estimated errors (assumed as 1 − σ variation), where the errors were assumed to follow a Gaussian distribution.

To compute the orbits, we adopt a mean RV of −325.95 ± 6.6 km s⁻¹ computed from the combined APOGEE-2 RVs of the two program stars and Holger Baumgardt's RV compilation ¹⁷ of 34 stars, which lie in the RV range between −341.13 and −310.22 km s⁻¹. From those stars, we select objects having a Renormalized Unit Weight Error below 1.4, as extracted from Gaia EDR3, which allows us to discard sources with problematic astrometric solutions (see, e.g., Lindegren et al. 2018). This reduces the 36 stars with RV information to 27 stars with both reliable proper motions and RV information, which were considered to compute the nominal proper motion of VVV CL001 after applying a 3 − σ clipping to the data, e.g., only 25 out of these 27 sources lie inside 3 − σ of the nominal proper motion of the cluster, as shown in panel (b) of Figure 1. From this procedure, the nominal proper motion of VVV CL001 is $({\mu }_{\alpha }\cos (\delta ),{\mu }_{\delta })=(-3.41,-1.97)$ mas yr⁻¹, with an assumed uncertainty of 0.5 mas yr⁻¹. The heliocentric distance (d_⊙) of VVV CL001 remains uncertain; for this reason we assume three possible estimates (5.5, 8.0, and 10.5 kpc) close to the best isochrone fit, as shown in panel (f) of Figure 1.

The main orbital elements and the ensemble of orbits of VVV CL001 are displayed in panels (a) to (c) and (d) of Figure 4. The orbital parameters reveal that VVV CL001 lies on a radial and highly eccentric (>0.8) halo-like orbit with rather small excursions above the Galactic plane (${Z}_{\max }\lt 3$ kpc), and pericentric (r_peri) distances below 1 kpc. We also find that VVV CL001 exhibits a retrograde and prograde sense at the same time when assuming a close heliocentric distance, while it is exclusively retrograde at and beyond the Bulge, similar to other GCs in the inner Galaxy (see Pérez-Villegas et al. 2020). Therefore, a more robust heliocentric distance estimation will better constrain these dynamics scenarios.

Zoom In Zoom Out Reset image size

Panel (d) of Figure 4 shows that the orbital energy configuration of VVV CL001 is comparable to that of Galactic GCs associated with the major accretion events such as Sequoia (Seq.) and GES (see, e.g., Myeong et al. 2018; Massari et al. 2019), indicating that VVV CL001 could be the fossil relic of one of these accreted dwarf galaxies.

Baumgardt & Hilker (2018) performed N-body simulations of star clusters, and found that they could reproduce the surface-density profile of VVV CL001, finding a present-day mass of 9 × 10⁴ M_⊙.

With the available RV data, we match the line-of-sight dispersion profiles to the N-body simulations of VVV CL001, as shown in panel (e) of Figure 4, and thus determine the most likely mass of the cluster from kinematics constraint. We adopted three radial bins (with bin centers of 6', 15', and 30'), chosen to ensure that at least 10 stars were in each bin; resulting in the three blue dots shown in panel (e) of Figure 4. With the new data in panel (e) of Figure 4, we find σ₀ ∼ 6.6 km s⁻¹. This yields a present-day estimated mass of ∼2.1 × 10⁵ M_⊙, which suggests that VVV CL001 is two times more massive than previously thought, and approximately four times more massive than ESO280-SC06 (the most metal-poor GC known in the MW).

We have performed the first near-IR high-resolution spectral analysis of two likely members of the GC VVV CL001, a cluster obscured by the heavy extinction and high field-star density in the direction of the Galactic Bulge. Based on high S/N APOGEE spectra, we measure a mean [Fe/H] metallicity of −2.45 ± 0.1, which makes VVV CL001 possibly the most metal-poor GC known, only slightly higher than ESO280-SC06 at −2.48.

VVV CL001 is very close in projection to UKS 1, which motivated Minniti et al. (2011) to suggest the possibility that they could be gravitationally bound. However, the RV of VVV CL001 is too large, and the orbits are very different, which allows us to rule out the binary-cluster scenario.

We find that the α-element abundances of VVV CL001 are typical of the lowest metallicity GCs known. Spectra for more members are required in order to confirm the presence of MPs in this cluster.

For the derivation of the age and reddening, we employed the new code SIRIUS (Souza et al. 2020). As shown in panel (e) of Figure 1, we derived a median age of $\sim {11.9}_{-4.05}^{+3.12}$ Gyr, indicating that the very metal-poor GC VVV CL001 is among the oldest and most massive (∼2.1 × 10⁵ M_⊙) MW clusters.

A dynamical analysis of VVV CL001 reveals that this object has a radial and highly eccentric halo-like orbit confined inside the Sun's galactocentric distance. Both its metallicity and orbit favor the interpretation of VVV CL001 being a GC that belongs to an early accretion event in the MW corresponding to the either the Seq or GES dwarf galaxies. Finally, multiband photometry in the near-IR will be useful to identify other stellar tracers, such as RR Lyrae stars (if any) toward VVV CL001, which will significantly help to constrain the distance, age, and origin of the cluster.

We thank the anonymous referee for helpful comments that greatly improved the paper. We warmly thank Holger Baumgardt for providing his published numerical N-body modeling of the line-of-sight velocity dispersion of VVV CL001. J.G.F.-T. is supported by FONDECYT No. 3180210. D.M. is supported by the BASAL Center for Astrophysics and Associated Technologies (CATA) through grant AFB 170002, and by project FONDECYT Regular No. 1170121. S.O.S. acknowledges the FAPESP PhD fellowship 2018/22044-3. T.C.B. acknowledges partial support for this work from grant PHY 14-30152: Physics Frontier Center/JINA Center for the Evolution of the Elements (JINA-CEE), awarded by the US National Science Foundation. D.G. gratefully acknowledges support from the Chilean Centro de Excelencia en Astrofísica y Tecnologías Afines (CATA) BASAL grant AFB-170002. D.G. also acknowledges financial support from the Dirección de Investigación y Desarrollo de la Universidad de La Serena through the Programa de Incentivo a la Investigación de Académicos (PIA-DIDULS). S.V. gratefully acknowledges the support provided by Fondecyt reg. n. 1170518. B.B. acknowledges partial financial support from FAPESP, CNPq, and CAPES—Finance Code 001. A.P.-V. and S.O.S. acknowledge the DGAPA-PAPIIT grant IG100319. L.H. gratefully acknowledges support provided by National Agency for Research and Development (ANID)/CONICYT-PFCHA/DOCTORADO NACIONAL/2017-21171231. A.R.-L. acknowledges financial support provided in Chile by Comisión Nacional de Investigación Científica y Tecnológica (CONICYT) through the FONDECYT project 1170476 and by the QUIMAL project 130001.

Funding for the Sloan Digital Sky Survey IV has been provided by the Alfred P. Sloan Foundation, the U.S. Department of Energy Office of Science, and the Participating Institutions. SDSS-IV acknowledges support and resources from the Center for High-Performance Computing at the University of Utah. The SDSS website is www.sdss.org.

SDSS-IV is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS Collaboration including the Brazilian Participation Group, the Carnegie Institution for Science, Carnegie Mellon University, the Chilean Participation Group, the French Participation Group, Harvard-Smithsonian Center for Astrophysics, Instituto de Astrofísica de Canarias, The Johns Hopkins University, Kavli Institute for the Physics and Mathematics of the Universe (IPMU)/University of Tokyo, Lawrence Berkeley National Laboratory, Leibniz Institut für Astrophysik Potsdam (AIP), Max-Planck-Institut für Astronomie (MPIA Heidelberg), Max-Planck-Institut für Astrophysik (MPA Garching), Max-Planck-Institut für Extraterrestrische Physik (MPE), National Astronomical Observatory of China, New Mexico State University, New York University, the University of Notre Dame, Observatório Nacional/MCTI, The Ohio State University, Pennsylvania State University, Shanghai Astronomical Observatory, United Kingdom Participation Group, Universidad Nacional Autónoma de México, University of Arizona, University of Colorado Boulder, University of Oxford, University of Portsmouth, University of Utah, University of Virginia, University of Washington, University of Wisconsin, Vanderbilt University, and Yale University.

This work has made use of data from the European Space Agency (ESA) mission Gaia (http://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, http://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. The Master of Astronomy of the Universidad de Antofagasta, who are associated as co-authors of this manuscript, thanks the Graduate School of the Universidad de Antofagasta for their support and for the "Beca de Excelencia" (Scholarship of Excellence). Simulations have been executed on HPC resources on the Cluster Supercomputer ATOCATL from Universidad Nacional Autónomade México (UNAM).