© ESO 2022

1. Introduction

One of the most important questions in astrophysics that still needs an explanation is how the Milky Way (MW) bulge was formed and how it has evolved to the present state (Di Matteo et al. 2015; Barbuy et al. 2018). Existing models (e.g., Kormendy & Kennicutt 2004; Gadotti 2009) suggest that bulges can be classified as classical bulges, if they result from a violent event (i.e. galaxy mergers or sinking of giant gas clumps), or disk-like bulges, which may be the consequence of internal processes (e.g., disk instabilities) on longer timescales. In the first case older stellar populations are expected within a spherical structure; in the second a wide range of ages for stellar populations are expected in a flattened structure.

Very useful tools for reconstructing the MW history and for characterising its current evolutionary stage are globular clusters (GCs). Being the first stellar associations (along with dwarf spheroidal galaxies) formed in the early Universe with typical lifetimes of 12–14 Gyr (Salaris & Weiss 2002), GCs retain fingerprints of past events that have occurred in their host galaxy (e.g., mergers and interactions). Thus, they provide a channel for tracing the properties of the MW, such as its structure, dynamics, compositions as well as the dark matter morphology. However, the number of known MW GCs seems to be small in comparison with the Andromeda galaxy (Harris et al. 2013; Minniti et al. 2017a) or the Galactic open clusters (OCs; Bonatto & Bica 2010; Cantat-Gaudin et al. 2018, 2020), indicating that our Galaxy has been inefficient at forming GCs, or that strong processes have destroyed these objects, or that many GCs have yet to be discovered. During the last two decades many new GCs have been detected in the Galactic halo thanks to optical photometric surveys like the Sloan Digital Sky Survey (SDSS; York et al. 2000), and also across the Galactic bulge and disk with surveys such as the near-infrared (NIR) Two Micron All-Sky Survey (2MASS, Skrutskie et al. 2006) and the VISTA Variables in the Via Lactea (VVV; Minniti et al. 2010; Saito et al. 2012) and its extension survey (VVVX; Minniti 2018). Already more than 300 star cluster candidates have been discovered within the VVV–VVVX images, but a dedicated analysis needs to be done in order to confirm their physical nature (e.g., Minniti et al. 2017a,b). The discovery of new GCs towards the inner regions of our Galaxy can be useful to constrain formation and evolution models of the MW, especially when accurate physical parameters are estimated. Significant effort has been devoted in recent years to studying the real nature of GC candidates, and indeed many of them have been confirmed. A few examples are VVV-CL001 (Minniti et al. 2011; Fernández-Trincado et al. 2021a), VVV-CL002, and VVV-CL003 (Moni Bidin et al. 2011); FSR 1716 (a.k.a. Minni 22; Minniti et al. 2017b); RLGC 1 and 2 (Ryu & Lee 2018); Camargo 1107, 1108, and 1109 (Camargo & Minniti 2019); FSR 1758 (Barbá et al. 2019; Romero-Colmenares et al. 2021); Garro 01 (Garro et al. 2020); UKS 1 (Fernández-Trincado et al. 2020); Patchick 99 (Garro et al. 2021b); FSR 19 and FSR 25 (Obasi et al. 2021); VVV-CL160 (Minniti et al. 2021a); Minni 48 (Minniti et al. 2021b); FSR 1776 (also known as Minni 23; Dias et al. 2022); Ton 1 (Fernández-Trincado et al. 2021b); and nine other GCs in Garro et al. (2022).

The present study provides the photometric analysis of 12 GC low-latitude candidates, shown in Fig. 1, including measurements of physical parameters, in order to reveal their real nature. In Sect. 2 we introduce these new GC candidates (see Table 1). In Sect. 3 we briefly quote the datasets adopted. In Sect. 4 we describe the methods used to investigate them and how we derive their main physical parameters. We provide individual notes about each of these clusters in Sect. 5. Finally, a summary and discussion are presented in Sect. 6.

Fig. 1. Galactic distribution of GCs towards the MW bulge. Star clusters investigated in this work are highlighted with red points, while open green squares represent the Galactic GCs by Vasiliev & Baumgardt (2021) in the Gaia EDR3 footprint.

2. Cluster candidates

The sample targets, listed in Table 1 and displayed in Fig. 1, include 12 GC candidates. These were selected mostly from the clusters discovered by members of the Deep Sky Hunters collaboration, a world-wide group of amateur astronomers dedicated to finding new deep sky objects such as open clusters and planetary nebulae (Kronberger et al. 2006, 2016). The new clusters were originally discovered through a visual inspection of the DECam Plane Survey (DECaPS; Schlafly et al. 2018) as concentrated local enhancements of the stellar density (see Fig. A.1). In addition to the recent DECaPS finds, we also include some older discoveries that have yet to be properly analysed: ESO 92-18 was discovered in the early 1980s as part of the ESO/Uppsala Survey catalogues (Lauberts 1982), but it also was detected by Froebrich et al. (2008) and was listed as MWSC 3250 by Kharchenko et al. (2013); Gaia 2 was an independent Digitized Sky Survey (DSS) discovered by Kronberger et al. (2012), but also detected by Koposov et al. (2017) from Gaia Data Release 2 (Gaia Collaboration 2018b) observations; Riddle 15 was discovered by Kronberger et al. (2006), and also listed as MWSC 3063 by Kharchenko et al. (2013). These clusters are analysed here for the first time, and we have kept their original names after their discoverers (Kronberger, Patchick, Riddle, and Ferrero).

3. Optical and NIR datasets

Investigating these 12 clusters appears a tricky task to perform as many of these objects are located in obscured regions or behind dust clouds, or they can be so faint or diffuse that it is difficult to distinguish them against the background. The presence of patchy differential reddening in front of some of these clusters represents an additional obstacle to their detailed analysis. Hence, it is necessary to employ distinct datasets at different wavelengths in order to overcome these complications and obtain trustworthy results. We especially rely on NIR datasets as in these bands the effect of extinction due to the dust is greatly reduced with respect to the optical, and since giant stars appear very luminous at these wavelengths, this can help to derive reliable physical parameters, such as distance and metallicity. A brief description of the available datasets is presented below.

3.1. VVV and VVVX photometry

We use the NIR datasets of the VVV and VVVX surveys. The data were obtained using the 4.1 m wide-field Visible and Infrared Survey Telescope for Astronomy (VISTA; Emerson & Sutherland 2010) at the ESO Paranal Observatory. The VISTA Telescope tile field of view is 1.501 deg², and 196 tiles are used to map the bulge area (−10° < l < +10°, −10° < b < +5°) and 152 tiles for the disk (−65° < l < −10°, −2.25° < b < +2.25°). The extended VVVX campaign is sampling a larger area of the sky with the goal of extending the footprint of the VVV survey by surrounding this previous area in the bulge and the mid-plane areas at longitudes from 295° to 10°. The VVV and VVVX photometric datasets are both divided into several bulge + disk tiles: d001 to d152 in the disk and b201 to b396 in the bulge for the VVV regions, whereas b401 to b512 in the bulge and e601 to e988 in the disk regions for the VVVX regions. The VVVX area also includes an extension from 230° to 295° longitude and −2° to +2° latitude, which are sampled in the tiles e1001 to e1180.

The VVV data was reduced at the Cambridge Astronomical Survey Unit (Irwin et al. 2004) and further processing and archiving was performed with the VISTA Data Flow System (Cross et al. 2012). For the VVV clusters, we extracted the photometry from the VVV catalogue published by Alonso-García et al. (2018). This catalogue was obtained using point spread function techniques, and is calibrated into the VISTA photometric system (González-Fernández et al. 2018). A preliminary catalogue from the VVVX data, obtained following a similar approach (Alonso-García et al., in prep.), was used to extract the photometry for the clusters in the VVVX footprint.

The advantage of using this public survey is that its passbands (ZYJHK_s in VVV and JHK_s in VVVX) allows us to study in detail evolved stars in the colour-magnitude diagram (CMD), in particular red clump (RC) stars, which are good metallicity and distance indicators. Unfortunately, not all star clusters considered here fall under the VVVX coverage. In the VVV–VVVX images we detected the following star clusters:

• Ferrero 54 in the VVVX tiles e706,705;

• Kronberger 100 in the VVVX tile e1032;

• Kronberger 143 in the VVV tile d003;

• Patchick 125 in the VVVX tile e932;

• Patchick 126 in the VVVX tile e676.

3.2. 2MASS photometry: brighter stars

The 2MASS is an all-sky survey in the J, H, and K_s NIR bands. We recovered this photometry for two basic reasons. For all clusters that have a VVV–VVVX counterpart, the 2MASS photometry allows us to build a CMD with stars from the main sequence (MS) to the red giant branch tip (RGB-tip), while VVV–VVVX stars with K_s < 11 mag are saturated. To do this, since the 2MASS and VISTA photometric systems are linked but different, we transformed the 2MASS photometry into the VISTA magnitude scale following González-Fernández et al. (2018). On the other hand, for the other clusters, such as Kronberger 99, Kronberger 119, Patchick 122, Riddle 15, FSR 190, ESO 92-18 and Gaia 2, which are outside the VVVX coverage, we obtained the NIR information for the brighter stars. In a few cases the 2MASS star sample appears to be small, although we are still able to establish some of the physical parameters of our targets, such as metallicity.

3.3. Gaia EDR3 photometry

The Gaia Early Data Release 3 (EDR3, Gaia Collaboration 2021) photometry is handled to build vector proper motion (VPM) diagrams, to separate star cluster members from background and foreground stars, and to confirm the cluster nature of our sample. Its precise astrometry is also used to eliminate all nearby stars with distances D < 2 kpc (Sect. 4.1) from our source catalogues. Additionally, this photometry gives the possibility to construct deeper CMDs for those clusters without VVV–VVVX data, as we can appreciate for ESO 92-18 and Gaia 2 CMDs (Fig. B.2). No photometric colour or magnitude cuts were applied to the Gaia EDR3 data.

4. Methods

4.1. Decontamination procedure

This analysis employs the combination of the VVV–VVVX and 2MASS photometry in the J, H, and K_s bands, and the optical Gaia EDR3 dataset in order to investigate the nature of these candidates.

First, we join the VVV + Gaia EDR3 and/or the 2MASS + Gaia EDR3 datasets using a 0.5″ matching radius. We treat the two NIR datasets separately, making a simple magnitude cut at K_s = 11 mag. After that we apply a decontamination procedure in order to build a clean catalogue with likely star members of every GC candidate, adopting the same procedure applied by Garro et al. (2020, 2021b). As a preliminary step we exclude all nearby stars with parallax values larger than 0.5 mas. Subsequently, we select stars within a given radius r from the cluster centre (see Table 1). We follow two methods in order to derive the cluster size: first, we inspect the density diagrams as a function of the sky position, as shown in Fig. 2, in order to visually identify the cluster dimensions as an overdensity; second, we apply the Gaussian kernel density estimation (KDE; e.g., Rosenblatt 1956; Parzen 1962). Briefly, considering the scatterplot on the left-hand side of Fig. 3, overlapping points make the figure hard to read. Even worse, it is impossible to determine how many data points are in each position. In this case a possible solution is to cut the plotting window into several bins (100–300 bins), and represent the number of data points in each bin by a colour. Following the shape of the bin, this makes a 2D histogram. It is then possible to obtain a smoother result using Gaussian KDE. Its representation is called a 2D density plot, and we add isodensity contours to denote each step. Therefore, we balance both the KDE contours and the visual overdensities in order to select the stars within a given radius (Table 1). We adopt a GCs radius of ∼3′ for Kronberger 99, Kronberger 143, Ferrero 54, ESO 92-18, and Gaia 2; a larger size (r ≈ 10.2′) for FSR 190; and smaller dimensions (r < 2.5′) for Kronberger 100, Kronberger 119, Patchick 122, Patchick 125, Patchick 126, and Riddle 15. At this level we benefit from exquisite Gaia PM measurements, constructing the vector PM diagrams (Fig. B.1), in order to keep only stars with a high probability of cluster membership. We are able to estimate the mean cluster PMs, listed in Table 1, using the sigma-clipping technique. In our final datasets we include only stars within 1 mas yr⁻¹ from the mean cluster PM values. Even though the assumption of a fixed PM radius value to select the final star sample is rigorous, it is linked to the fact that the PM overdensities, shown in the VPM diagrams (Fig. B.1), which indicate the presence of a cluster, are very small and peaked in most cases. This allowed us to keep only stars that are likely members of the clusters (with >50% probability), minimising the likelihood of including outliers, and of excluding member stars. For this reason we also calculated the PM probability for each star within the cluster size, as shown in Fig. B.1. However, a very small fraction of contaminants in the final catalogues may be expected since these regions are characterised by high stellar crowding and by disk stars with similar PMs, as demonstrated by Garro et al. (2021b). Even if this occurs, it does not have a significant effect on our final photometric analysis as the overdensity peaks are clear in the VPM diagrams (Fig. B.1) against the background noise, ensuring that the final catalogues include likely cluster members.

Fig. 2. Gaia EDR3 density maps, cleaned from nearby stars, centred on the ESO 92-18 (left) and Kronberger 100 (right) fields. These two clusters are used to visually show the different target sizes. The greener areas are higher densities, while the bluer areas are lower densities.

Fig. 3. KDE technique applied to select the likely size of a cluster during the decontamination procedure. From left to right: scatterplot, 2D histogram, Gaussian KDE, and 2D density with contours (or iso-densities) are shown for the ESO 92-18 (top panels) and Kronberger 100 (bottom panels) fields, used as examples. The green and yellow areas are indicate higher densities, while the blue areas show lower densities.

The resulting values are listed in Table 1, where we summarise the cluster identification (ID), the position, the radius cuts, the number of star members (N_⋆), and the mean cluster PMs both in RA (μ_α* = μ_α × cos δ) and in Dec (μ_δ).

4.2. Estimation of physical parameters

Once the final catalogues were obtained, we derived the main physical parameters for each candidate: reddenings and extinctions, distances, total luminosities, ages, and metallicities. The derivation of these parameters is challenging because in those areas crowded with additional dust and gas (see Fig. 4) the observations and precise measurements are more complicated to perform. Moreover, towards the innermost regions the differential reddening becomes more prominent, increasing the uncertainties in the determination of distances. Then, the extinction and distances are estimated following the same procedures described in our previous works (Garro et al. 2021b; Minniti et al. 2021b). We adopt existing reddening maps, such as those by Gonzalez et al. (2011), Soto et al. (2019), and Surot et al. (2019) in the NIR, and by Schlafly & Finkbeiner (2011) in the optical passbands. Therefore, we use the following relations to derive the colour excess and the extinction: A_V = 3.1 × E(B − V), A_Ks = 0.11 × A_V, A_G = 0.86 × A_V, A_Ks = 0.75 × E(J − Ks), A_G = 10.115 × A_Ks, and A_G = 2.0 × E(BP − RP). We find extinction spans of 0.09 ≲ A_Ks ≲ 0.86 mag and 0.89 ≲ A_G ≲ 4.72 mag. The extinction values are mandatory to compute reliable heliocentric distances, using the position of the red clump (RC). The advantage of using $M_{\mathit{Ks}}^{\mathrm{RC}}$ arises from its simple identification and low sensitivity to extinction by interstellar dust (e.g., Salaris & Girardi 2002). Therefore, we adopt the intrinsic RC magnitude $M_{\mathit{Ks}}^{\mathrm{RC}}=-1.606\pm 0.009$ mag and $M_G^{\mathrm{RC}}=0.383\pm 0.009$ mag from Ruiz-Dern et al. (2018), the observed NIR and optical RC magnitude listed in Table 2 (Cols. 2 and 3), and the extinction values (Cols. 3 and 4) in order to derive the distance moduli (Col. 6). These parameters can be translated into heliocentric distances (Cols. 7 and 8), finding 4.4 ≲ D_NIR ≲ 18.1 kpc from the NIR photometry. The NIR and optical distance values agree well within the errors since we find 3.7 ≲ D_GEDR3 ≲ 19.5 kpc from the Gaia photometry. However, clusters with blue horizontal branches usually do not show a defined RC, especially in poorly populated CMDs. Therefore, we derive the colour excess for the optical passband comparing the cluster horizontal branch (HB) and RGB colours with the absolute colours by Gaia Collaboration (2018a).

Fig. 4. Optical DECaPS (left) and NIR VVVX (right) images of the sky region centred on Ferrero 54. The field of view is 200″ × 200″, oriented along the Galactic coordinates, with Galactic longitude increasing to the left and Galactic latitude increasing to the top. Veils of gas and dust are evident as green filaments in the optical passband. They are part of the Vela supernova remnant complex.

The extinction coefficient that we use in our analysis is R_V = 3.1, which is the standard value for diffuse interstellar medium (e.g., Sneden et al. 1978). Recent works (Nataf et al. 2013; Pallanca et al. 2021a,b; Souza et al. 2021) have demonstrated that a low R_V ≈ 2.5−2.7 is more appropriate to reproduce the bulge populations, especially in the regions where Patchick 125 and 126 are located. Clearly, a different extinction value has an effect on the colour excess and distance estimates. We find that the distances increase by ∼0.1 kpc for Patchick 125 and ∼0.16 kpc for Patchick 126 when a low R_V is adopted. Nonetheless, the two distance values, obtained adopting R_V = 3.1 and 2.5, are in good agreement within the errors. Additionally, we note that when we use R_V = 2.5, and we keep the metallicities and ages as listed in Table 2, the isochrones do not fit the points properly (especially along the RGB sequences) in the CMD unless a higher metallicity is used, as shown in Fig. 5. However, we argue that R_V = 3.1 is a good compromise if we use the Patchick 125 metallicity of [Fe/H] = −1.70, spectroscopically found by Fernández-Trincado et al. (2022) (see Sect. 5.6).

Fig. 5. Patchick 125 (bottom panels) and Patchick 126 (top panels) NIR CMDs. In the left panels we compare two isochrones using two different extinction coefficients: RV = 3.1 (yellow points) and RV = 2.5 (red points). In the right panels we fix the RV = 2.5 and the ages as listed in Table 2, and we change the metallicity as specified in the legend.

Moreover, we derive their galactocentric X, Y, Z coordinates in kiloparsec using the NIR distances. As shown by Fig. 6, we place our candidates at their respective distances R_G from the Galactic centre, assuming R_⊙ = 8.2 kpc (GRAVITY Collaboration 2019). We find that they are located at galactocentric distances ranging between 2.94 kpc (for Patchick 126) and 14.06 kpc (for Riddle 15), as listed in Table 2. Additionally, we calculate the distance above or below the Galactic plane, using the relation Z = D × sin(b) and assuming Z_⊙ = 0 kpc, spanning −1.23 ≲ Z ≲ 0.78 kpc.

Fig. 6. Galactocentric distribution of star clusters analysed in the present work (red points) and Galactic GCs (green open squares) by Baumgardt & Vasiliev (2021). The black cross represents the position of the Galactic centre, while the blue star is the position of the Sun.

Once distances and mean PMs are obtained, we can compute their tangential velocities ($V_T^{\mathrm{RA}}$ and $V_T^{\mathrm{Dec}}$ ). Moreover, we compare the PMs and V_T distributions with those of known and well-characterised MW GCs by Vasiliev & Baumgardt (2021), as shown in Fig. 7. We obtain negative μ_δ for Riddle 15, FSR 190, and Patchick 126, which show −8.0 < μ_δ < 0 mas yr⁻¹ and −10.0 < μ_α* < 0 mas yr⁻¹, while all the others display μ_δ > 0 mas yr⁻¹ and the same μ_α* range. In addition, for V_T we clearly obtain a trend similar to that of the PMs since we find negative $V_T^{\mathrm{Dec}}$ for Riddle 15, FSR 190, and Patchick 126, which are $-300<V_T^{\mathrm{Dec}}<-100$ km s⁻¹, while for the other clusters $V_T^{\mathrm{Dec}}$ is always positive between 0 and 200 km s⁻¹, while the $-300<V_T^{\mathrm{RA}}<0$ km s⁻¹ for all clusters. The clusters exhibit the longitude dependence due to the solar motion around the Galactic centre, but as a whole the new GCs identified here are consistent with the bulk motions observed for the rest of the MW globular cluster system.

Fig. 7. Gaia EDR3 PM distribution of star clusters in the present work (red points) and Galactic GCs (open green squares) by Vasiliev & Baumgardt (2021).

Reddenings and distances are crucial to deriving the cluster metallicities and ages. One of the methods that allows us to determine these parameters with reasonable reliability is the isochrone-fitting method. Our CMDs in Fig. B.2 are populated by evolved stars, such as RC, blue horizontal branch (BHB), and red giant branch (RGB) stars. These stars are excellent indicators of the global metallicity content of a stellar population as they are more sensitive to metallicity variations (McQuinn et al. 2019). However, for Kronberger 100, ESO 92-18, and Gaia 2, fainter main sequence stars are detected by Gaia EDR3 photometry. We use the PARSEC models (Bressan et al. 2012; Marigo et al. 2017) in order to obtain the metallicity for each cluster, finding a range of −1.8≲ [Fe/H] ≲ + 0.3 with a mean error of ±0.2 dex. On the other hand, it is well known that the main sequence turn-off (MSTO) is the best point to derive the absolute age of a cluster. Even so, some of our Gaia and VVVX–2MASS CMDs are not deep enough to reach the MSTO point (or go to ∼2 mag below), as we can see in Fig. B.2, for example, for Riddle 15, Patchick 126, and Kronberger 119. Therefore, our determination of the ages of these clusters is approximate, and deeper observations are necessary to confirm our results. However, we find a range of 2< age < 14 Gyr, in close agreement with old and intermediate-age Galactic GCs. More precise ages are evaluated for Kronberger 100, Kronberger 119, Kronberger 143, Patchick 125, ESO 92-18, and Gaia 2, as we detail in Sect. 5. Additionally, we determine the errors in age and in metallicity by testing isochrones at a fixed metallicity and a fixed age, respectively, until they no longer fit the cluster sequence simultaneously in the NIR and optical CMDs (see Fig. B.2).

On the other hand, we apply the χ² analysis to derive the age and the metallicity as an independent method. We perform this analysis for all clusters on the Gaia CMDs since in many cases optical CMDs are deeper than NIR CMDs, using a family of PARSEC isochrones. We adopt ages between 1 and 13 Gyr, and we change the metallicity values, depending on the [Fe/H] estimates, listed in Table 2. Figure 8 shows the χ² values as a function of age, whereas we list the likely age and metallicity corresponding to the minimum ${\chi }_{min}^2$ value in Table 3. We obtain a good agreement between the metallicity estimates obtained with the isochrone fitting and χ² methods, except for Kronberger 99 for which the χ² method yields a lower metallicity ([Fe/H] = − 1.4). Nonetheless, the age values represent a lower limit again, especially for those clusters with few points, such as Patchick 122. This was expected since neither the MS nor the MSTO are reached in some CMDs. For ESO 92-18 (see Fig. 8, right panel) and Gaia 2 we obtain an excellent agreement (within 1σ) between the two methods because the MSTO and the higher part of the MS are above the detection limit. Conversely, the FSR 190 CMD is very noisy; therefore, our procedure does not show a clear minimum for the age. For this reason we suggest that the estimates for that cluster found by the previous method and by the comparison with the literature (see Sect. 5.9) are more reliable.

Fig. 8. χ2 estimate as a function of age for three representative clusters: ESO 92-18, Patchick 125, and Patchick 126. The colour of the lines depends on the metallicity, specified in the legend. The black dashed lines indicate the minimum χ2 value at the relative age. The age and metallicity that correspond to the ${\chi }_{min}^2$ are listed in Table 3. Right panel: Gaia EDR3 CMD for ESO 92-18. The PARSEC isochrone (black dotted line) was fit using the age and metallicity obtained by the χ2 analysis and listed in Table 3. The reddening, distance modulus, and extinction are listed in Table 2.

Furthermore, it is important to note that the PARSEC isochrones are available for [α/Fe] = 0, a typical value associated with OCs and metal-rich GCs. In the case of the metal-poor clusters, such as Kronberger 99, Kronberger 143, and ESO 92-18, an enhanced [α/Fe] value should be adopted. Therefore, for these clusters we also used the Dartmouth Stellar Evolution Database (Dotter et al. 2008), adopting enhanced [α/Fe] values, and keeping age and metallicity fixed, as listed in Table 2. In Fig. 9 we compare the Dartmouth and PARSEC isochrones, finding a good agreement between these models. However, we suggest that these clusters should have [α/Fe]= + 0.2.

Fig. 9. Gaia EDR3 optical CMDs for the clusters: Kronberger 99 (left panel), Kronberger 143 (middle panel), and ESO 92-18 (right panel). The green, yellow, and pink lines are the Dartmouth isochrones with [α/Fe]= + 0.4, +0.2, 0, respectively. The dotted red lines are the PARSEC isochrones. We keep the age and metallicity fixed as listed in Table 2.

We categorise all clusters with age ≳ 8 Gyr and metallicity [Fe/H] ≲ − 0.1 dex as GCs, and the rest as OCs.

Finally, the total luminosity in the K_s band for each cluster is derived following the same approach used in Garro et al. (2021a). We first derive the absolute magnitude (M_Ks) measuring the cluster total flux. Although Kharchenko et al. (2016) have demonstrated that the larger contribution in luminosity arises from the brightest stars since the first 10–12 members accumulate more than half of the integrated luminosity of the cluster, we are aware that the resulting M_Ks are a lower limit because our photometry does not include fainter and lower mass stars. Therefore, in order to quantify the fraction of luminosity missed we adopt the same procedure used in Garro et al. (2021a), comparing the luminosity of our sample with those of known Galactic GCs. Once this luminosity fraction is obtained, we add this to the absolute magnitude and convert it into the absolute magnitude in V band. In this way we obtain an empirical correction to our clusters’ luminosities under the assumption of similarity among Galactic GCs. We find that all clusters have low luminosities with −3.5 < M_V < −6.9 mag, much fainter than the peak of the MW GC luminosity function (GCLF), which is $M_V^{\mathrm{MW}}=-7.4\pm 0.2$ mag from Harris (1991), Ashman & Zepf (1998), also in agreement with the value found by Garro et al. (2021a) of $M_V^{\mathrm{MW}}=-7.46\pm 1.04$ mag.

5. Fundamental parameters for each cluster

5.1. Kronberger 99

We analyse Kronberger 99 by matching Gaia EDR3 and 2MASS datasets because this cluster is outside of the VVVX field. Kronberger 99 appears as an overdensity both in the Gaia EDR3 and 2MASS images, especially when foreground stars are excluded. Inspecting the VPM diagram, we can easily extract the mean cluster PM, as shown in Fig. B.2; however, after the PM decontamination procedure, we find a very poorly populated cluster with 21 giant members. We select it as a GC candidate since a scarcely populated RC can be observed at K = 12.6 mag and G = 17.9 mag. The analysis yields an extinction and distance modulus for the cluster of A_Ks = 0.62 mag and (m − M)₀  =  13.23 mag, equivalent to D_NIR = 4.42 kpc, in good agreement with D_GEDR3 = 3.7 kpc, within 1σ. This also places this cluster at R_G = 9.42 kpc from the Galactic centre. The best PARSEC isochrone fit implies a metallicity of [Fe/H] = − 0.8 ± 0.2. On the other hand, due to the sparse population and lack of faint stars, it is very difficult to establish the age of this cluster. However, at [Fe/H] = − 0.8 we superimpose several isochrones with different age values, finding that all models with age < 6 Gyr deviate considerably from the position of the giant stars in the CMD, in particular the RGB sequence, which is not the case for older ages. In addition, computing the difference in magnitude between the HB and MSTO (e.g., Salaris & Cassisi 2005; Cassisi & Salaris 2013) we find that ΔK (HB-MSTO) is greater than 1.5 mag, which can be translated into a lower age limit of ∼6 Gyr, which is also in agreement with that found by the χ² analysis.

5.2. Kronberger 100

Kronberger 100 is investigated using the matches between the Gaia EDR3 + 2MASS and Gaia EDR3 + VVVX datasets. Approximately 20 bright stars are evolved off the MS, whereas the bulk of the probable cluster members form a populated MS, as we can appreciate from the CMDs (Fig. B.2). We estimate an extinction of A_Ks = 0.37 mag in the NIR and A_G = 2.76 mag in the optical. This places Kronberger 100 at a distance modulus of (m − M)₀ = 14.14 mag, equivalent to D_NIR = 6.76 kpc, which is in very good agreement with the heliocentric distance obtained from the Gaia EDR3 photometry (D_GEDR3 = 6.65 kpc.) Additionally, we obtain a distance from the Galactic centre of ∼10.0 kpc. The MSTO point lies near K_s = 15.5 ± 0.1 mag. The best PARSEC isochrone fit and the χ² analysis yield a mean cluster age of ∼3 Gyr; however, ages between 2 and 5 Gyr may be adequate to reproduce the sequences of this cluster, while the mean metallicity is [Fe/H] = + 0.3 ± 0.3 dex. Hence, we classify it as an open cluster (OC) containing 187 stars.

5.3. Kronberger 119

Matching the Gaia EDR3 and 2MASS catalogues, we analyse Kron-berger 119, finding a cluster with 182 members. Both in the 2MASS and Gaia CMDs a clear overdensity depicting the RC is seen at K = 13.9 mag and G = 17.4 mag, respectively. This helps us to derive the extinction of A_Ks = 0.24 mag and A_G = 1.85 mag, and also the distance modulus of (m − M)₀ = 15.27 mag, corresponding to D_NIR = 11.30 kpc. However, we found a smaller distance of D_GEDR3 = 9.8 kpc from the Gaia photometry. Adopting the NIR distance, we place this cluster at ∼11.54 kpc from the Galactic centre. As we can see in Fig. B.2, the MSTO is not reached in the 2MASS CMD; however, we have a deeper optical CMD, which allows us to derive a good age estimate since we evaluated the overdensity at G = 19.9 ± 0.1 mag as a MSTO point, evident also when the cluster luminosity function is used. Therefore, fitting the PARSEC isochores, we categorise Kronberger 119 as a young GC or an old OC with an age = 6 ± 1 Gyr and mean metallicity of [Fe/H] = − 0.5 ± 0.2 dex. However, we find a slightly younger age of 4.4 Gyr and lower metallicity of −0.9 with the χ² analysis.

5.4. Kronberger 143

Kronberger 143 falls in the VVVX coverage area, so we examine this cluster using a match of Gaia EDR3 + 2MASS and Gaia EDR3 + VVVX catalogues. From our decontamination procedure, we find that this cluster contains 148 members. In addition, well-defined RGB and RC regions, and the TO region are demarcated both in the NIR and optical CMDs (Fig. B.2). We estimate its main cluster parameters: extinction of A_Ks = 0.45 mag and A_G  =  3.12 mag, and (m − M)₀  =  14.06, equivalent to D_NIR  =  6.5 kpc, in excellent agreement with D_GEDR3  =  6.3 kpc. This places it at 7.81 kpc from the Galactic centre. We classify Kronberger 143 as a young GC with an age > 5 Gyr, which represents a solid lower limit; when younger isochrones are fitted, they do not reproduce any sequences (especially the MSTO region), while models with ages younger than 8 Gyr are more suitable for this cluster. Similar results are achieved through the χ² analysis.

5.5. Patchick 122

Patchick 122 is investigated using a combination of Gaia EDR3 and 2MASS photometry. This is a small (r ≈ 1.6′) and poorly populated cluster with 28 star members. We select it as GC candidate since a clear overdensity is obvious in the VPM diagram (Fig. B.2), but also a minute RC can be identified at K_s = 12.75 mag and G = 18.2 mag. We can speculate about it, suggesting that this may be a dissolved cluster or a survivor of strong dynamic processes, or that our data is incomplete. Nonetheless, we derive its main physical parameters, such as an extinction of A_Ks = 0.63 mag and A_G = 4.4 mag, and a distance modulus of (m − M)₀ = 13.72, corresponding to D_NIR = 5.6 kpc, in agreement within 2σ with the optical estimate of D_GEDR3 = 4.6 kpc. Using the NIR heliocentric distance, we place this cluster at R_G = 9.41 kpc from the Galactic centre. Patchick 122 may be a good GC candidate since when PARSEC isochrones are fitted we obtained an age > 6 Gyr and metallicity of [Fe/H] = − 0.5 ± 0.2 dex. Clearly, it is very difficult to reliably pinpoint the age of this cluster given the scarcity of stars detected. Hence, in order to overcome the age–metallicity degeneration, we estimate the age based on the vertical method. We find ΔG (HB-MSTO) ≳2.0 mag and ΔK (HB-MSTO) ≳2.5 mag, which can be translated into a lower limit of ∼6 Gyr. On the other hand, the χ² analysis suggests a younger age of 4.3 Gyr. For this reason, we fit models with age < 6 Gyr but keeping the metallicity fixed at [Fe/H] = −0.5 (obtained also by the χ² procedure), finding that the isochrones shift to lower magnitudes and towards the reddest part of the CMD, deviating significantly from the brightest stars. Therefore, we can exclude that this cluster is younger than 6 Gyr.

5.6. Patchick 125

We analyse Patchick 125 matching the Gaia EDR3 + 2MASS and Gaia EDR3 + VVVX datasets. After the decontamination procedure we find a small cluster with 146 star members. We derive its main physical parameters, obtaining an extinction of A_Ks = 0.33 mag in NIR and A_G = 2.56 mag in the optical, a distance modulus of (m − M)₀ = 15.18 mag, equivalent to a heliocentric distance of D_NIR = 10.90 kpc, in good agreement with the optical passbands (D_GEDR3 = 11.2 kpc). The NIR distance places this cluster at 3.23 kpc from the Galactic centre. Analysing the PM decontaminated CMDs, we can clearly observe an extended BHB, suggesting a metal-poor content. The best PARSEC isochrones fit yields an age = 14 ± 1 Gyr and [Fe/H] = − 1.80 ± 0.2 dex, confirming that Patchick 125 is an old GC. We evaluated its age using two observables: (i) calculating the difference in magnitude between the HB and MSTO, finding ΔK_s(HB-MSTO) > 4 mag and ΔG(HB-MSTO) > 3 mag, translating it into a solid limit of ∼13 Gyr; (ii) considereding the RGB-tip, which is less sensitive to age changes, but fitting models with age < 13 Gyr the isochrone-tip moves to brighter magnitudes. Therefore, we suggest that its age may be very old, in contrast with the age determind via the χ² method.

Additionally, we find that Patchick 125 is located within ∼10.2″ from the centre of the recently discovered GC Gran 3 (Gran et al. 2022). Therefore, we believe that the two objects are the same cluster. However, we keep the original name, which refers to the astronomer who discovered it first (Dana Patchick in 2018, internal communication). Comparing the present work with Gran et al. (2022), we find similar results. In particular, the mean cluster PMs μ_α* = −3.78 mas yr⁻¹ and μ_δ = +0.66 mas yr⁻¹, the Gaia EDR3 extinction A_G = 2.60 mag and the absolute magnitude in V band M_V = −6.02 mag are in excellent agreement with our results. However, they found a more metal-poor cluster with [Fe/H] = − 2.37 ± 0.18, relying on the relation between CaT EW and the magnitude of the HB in the Jonhson V filter. On the other hand, our metallicity is in good agreement with that derived by Fernández-Trincado et al. (2022) since they obtained a mean [Fe/H] = − 1.70, measuring the Fe I lines from the APOGEE-2S data.

Moreover, we searched for RR Lyrae stars in the cluster field. According to Clementini et al. (2019), there are 28 RR Lyrae located within 20′ from the centre of the cluster Patchick 125. Figure 10 shows their positions, magnitudes, and period versus amplitude diagrams. Based on their positions, magnitudes, and PMs, we consider that two of them are bona fide cluster members: Gaia DR2 5977224553266268928 and DR2 5977223144516980608, located at 79″ and 81″ from the cluster centre, respectively. These stars are shown as blue points in Fig. 10, and their mean magnitudes are G = 18.023 ± 0.016 and 17.753 ± 0.013 mag, and their mean PMs are μ_α* = −4.067 ± 0.330, μ_δ = 0.362 ± 0.224 mas yr⁻¹ and μ_α* = −3.793 ± 0.270, μ_δ = 0.853 ± 0.185 mas yr⁻¹, respectively. Their mean periods are P = 0.601940 and P = 0.738296 day, and their amplitudes are A = 0.872 and A = 0.633 mag in the optical passband (Clementini et al. 2019), respectively. The other RR Lyrae in the field, located beyond 2′ from the cluster centre may not be securely considered cluster members because of their different magnitudes and PMs. In particular, we exclude the other two variables located within 500 arcsec, because their PMs do not match the mean cluster PMs. Additionally, we derive their distances using the period-luminosity-metallicity (PL_KZ) relation by Alonso-García et al. (2021). First, we transformed the Patchick 125 metallicity (listed in Table 2) into log Z using the relation shown in Navarrete et al. (2017), assuming f = 10^[α/Fe] = 3 and Z_⊙ = 0.017. We find D = 11.06 ± 0.30 kpc for Gaia DR2 5977224553266268928 and D = 12.17 ± 0.50 kpc for Gaia DR2 5977223144516980608, both values in excellent agreement with the cluster distances obtained from the VVVX and Gaia EDR3 photometries. We also changed the α-element enhancement factor f = 1.58 (adopting the [α/Fe]= + 0.2), and we find for Gaia DR2 5977224553266268928 a distance of 10.4 kpc and for Gaia DR2 5977223144516980608 a distance of 11.5 kpc, in excellent agreement with the NIR cluster distance. However, we test our estimates using different PL_KZ relations by Muraveva et al. (2015), Gaia Collaboration (2017), and Navarrete et al. (2017), yielding slightly larger distance values. However, these relations are still within the mean errors of 0.3 and 0.5 kpc obtained by the comparison between these four PL_KZ relations. The presence of two RRab members of this cluster also support the conclusion that Patchick 125 is an old and metal-poor GC. In particular, the positions of these two member RR Lyrae in the period-amplitude diagram (shown in the bottom right panel of Fig. 10) suggest that Patchick 125 is an Oosterhoff type II GC. The mean ridge lines, used in Fig. 10, for the Oosterhoff type I and II populations are taken from Clement & Shelton (1999), assuming A_V/A_I ≈ 1.4 mag.

Fig. 10. Position (top left panel), magnitude-separation (top right panel), Gaia EDR3 PM-distribution (bottom left panel), and period-amplitude (bottom right panel) diagrams for the RR Lyrae stars located within 20′ from the centre of the cluster Patchick 125. The mean ridge lines for the Oosterhoff type I and II populations (left and right dotted lines, respectively) are from Clement & Shelton (1999). The two RR Lyrae members are shown as blue points.

5.7. Patchick 126

We use a combination of Gaia EDR3 and the NIR 2MASS and VVVX datasets in order to investigate Patchick 126. For this cluster, the CMDs convey a well-defined RGB and a poorly populated RC at K = 13.5 mag and G = 18.0 mag. We estimate an extinction of 0.441 mag in the NIR and 2.66 mag in the optical. From these parameters we obtain a distance modulus of (m − M)₀ = 14.66 mag, which places it at a heliocentric distance of 8.6 kpc, in agreement with D_GEDR3 = 7.4 kpc found from the Gaia photometry, within 2σ. We also derive its distance from the Galactic centre, R_G = 2.94 kpc. Our photometric analysis suggests that this may be a GC. We found the mean metallicity of the cluster is [Fe/H] = − 0.7 ± 0.3 dex and an age older than 8 Gyr, which were evaluated using the isochones fitting method. Although the metallicities derived from the two methods are in agreement, we find a younger age of 5.6 Gyr with the χ² procedure. Therefore, as done for the other clusters, we calculated the difference in magnitude between the HB and MSTO also for Patchick 126, obtaining ΔK_s(HB-MSTO) > 2.5 mag and ΔG(HB-MSTO) > 2.5 mag, which can be translated into a lower age limit of 8 Gyr. Additionally, taking its relatively low metallicity into account, Patchick 126 could be a younger GC, based on the 108 stars selected as members.

5.8. Riddle 15

Riddle 15 is studied using the match between Gaia EDR3 and 2MASS photometry because it is outside of the VVVX footprint. It appears confined within a radius smaller than 1′, which makes this candidate one of the smallest in our sample, suggesting that it may be located at a large distance. We find D_NIR = 18.1 kpc in NIR and D_GEDR3 = 19.5 kpc in the optical, and its galactocentric distance is R_G = 14.06 kpc. The main distance error source is related to the dust effect since we find higher extinction values of A_Ks = 0.52 mag and A_G = 3.09 in NIR and optical passbands, respectively. The only information that CMDs display is relative to its 83 brighter stars. Fortunately, we are able to reach below the HB, revealing an extended HB. We derive its metallicity using the PARSEC isochrones, yielding [Fe/H] = − 1.4 ± 0.2 dex. Taking into account both the low metal content and the magnitude level of the HB at G ≈ 20.0 mag, we believe that this cluster may be older than 10 Gyr, with the best isochrone fit suggesting an age of ∼13 Gyr, in good agreement with the χ² analysis. However, Bica et al. (2019) classify Riddle 15 as an OC candidate or a compact OC remnant candidate. They based their analysis on 2MASS data whose depth and angular resolution are clearly not appropriate for getting reasonable results for this faint, small, and distant cluster.

For this reason, we searched for RR Lyrae stars in the cluster field since they are good tracers of old and metal-poor populations. We found one RR Lyra type-ab variable star located at 43″ from the Riddle 15 cluster centre. This star is Gaia DR2 4320480608680062080, and has mean magnitude G = 20.254 ± 0.012, period P = 0.608333 day, epoch (JD) = 2456956.0055, amplitude A = 0.57 mag in the optical passband, and proper motions μ_α* = −1.90 ± 1.14 mas yr⁻¹ and μ_δ = −1.79 ± 1.05 mas yr⁻¹ (Clementini et al. 2019). In addition, we derive its absolute magnitude in G band adopting the M_G−[Fe/H] relation by Muraveva et al. (2018), assuming the [Fe/H] = − 1.4 as Riddle 15. Hence, this allows us to derive its distance D = 20.0 ± 0.5 kpc, in agreement with the cluster distance found from Gaia EDR3 photometry.

The position, magnitude, and PMs for this RR Lyra are consistent with cluster membership, also suggesting that Riddle 15 is a relatively old and metal-poor GC, excluding the OC nature.

There is another RR Lyra type-c candidate, Gaia DR2 4320478237881151488, located at 10′ from the Riddle 15 centre. However, this is likely a field star because it is too bright (G = 15.947 ± 0.006 mag) and its PMs are too large (μ_α* = 3.54 ± 0.07 mas yr⁻¹ and μ_δ = −13.79 ± 0.07 mas yr⁻¹) to be considered a cluster member.

5.9. FSR 190

It is very hard to analyse the field of FSR 190 cluster because of heavy crowding and differential reddening. This cluster appears to be highly reddened as we can see from its CMDs (Fig. B.2). We derived an extinction of A_Ks = 0.86 mag in NIR and A_G = 4.72 mag in optical. Additionally, the main information from its CMDs is related to 52 brighter star members, determined using a combination of Gaia EDR3 and 2MASS catalogues. From our analysis, we find that FSR 190 may be a GC; it is located at a heliocentric distance of D_NIR = 10.67 kpc and at a galactocentric distance of R_G = 11.11 kpc. In the optical passband we find a slightly smaller distance of D_GEDR3 = 9.5 kpc, but this value is more uncertain because of the high extinction. The best PARSEC isochrone fit suggests a metallicity of [Fe/H] = − 0.9 ± 0.2 dex. With the metallicity fixed, we searched for the best age, finding age > 9 Gyr. However, placing a small RC at K = 14.4 mag and G = 20.1 mag, and supposing that all the stars below K = 16.0 mag are MSTO stars, the cluster could have an age younger than 7 Gyr, but still in excess of 2 Gyr. if we use a family of isochrones with age < 7 Gyr, these differ considerably from the observational points of both CMDs. Therefore, we constrained the age of the cluster to be larger than 7 Gyr.

Although the analysis of this cluster turns out to be tricky and with many question marks, we are in full agreement with the results of Froebrich et al. (2008). They derived for FSR 190 similar values, such as metallicity ([Fe/H] −0.9 ± 0.4), age (age > 7 Gyr), and extinction (A_K = 0.8 ± 0.1 mag), and similar values of distances (D = 10.0 ± 1.0 kpc and R_G = 10.5 ± 0.8 kpc, adopting a R_⊙ = 7.2 kpc) within the errors. Our results also agree with the recent results of Cantat-Gaudin et al. (2018). Bica et al. (2019) also suggested that FSR 190 may be an OC or GC candidate, commenting that it could be a Palomar-like GC.

5.10. Ferrero 54

The main obstacle for the study of Ferrero 54 is the presence of a nebulosity of dust and gas in front of this cluster, as shown in the optical DECaPS image (Fig. 4), which disappears in the NIR image. These wisps are part of the Vela Supernova remnant complex. The decontamination procedure reveals a cluster including 122 giant stars, with a sparsely populated RC at K  =  13.25 mag and G = 17.8 mag. Its extinction is A_Ks = 0.60 mag and A_G = 3.0 mag. Using these values and the position of RC stars, we recover the distance modulus of (m − M)₀ = 14.27 mag, equivalent to D_NIR = 7.1 kpc, which is in excellent agreement with that found from the Gaia photometry of D_GEDR3 = 7.26 kpc. The NIR distance places Ferrero 54 at R_G = 11.5 kpc from the Galactic centre. The PARSEC isochrones fit yields a metallicity of [Fe/H] = − 0.2 ± 0.2 dex and an age ≈11 Gyr. We find that the metallicity estimates derived from the isochrone-fitting and χ² methods are in good agreement; on the other hand, the χ² procedure yields a younger age of 6.5 Gyr. Nevertheless, we give a solid lower age limit using the vertical method as done for other clusters, finding ΔK_s(HB-MSTO) > 3.0 mag and ΔG(HB-MSTO) > 3.0 mag, which can be translated into a lower limit of ∼10 Gyr. Additionally, GCs with almost solar metallicities are very scarce, and also in general they are located in the Galactic bulge. So if the high metallicity of Ferrero 54 is confirmed, it would make this an exceptional object among the group of Galactic GCs. Finally, based on our analysis, we confirm the GC nature for Ferrero 54.

5.11. ESO 92-18

As expected from the density diagrams (Figs. 2 and 3), ESO 92-18 is the most populated cluster here examined, containing 557 star members after the Gaia EDR3 PM decontamination procedure. We use the matched Gaia EDR3 + 2MASS datasets in order to have the NIR counterpart, even if we lose many fainter stars in the 2MASS CMD. However, reliable results can be achieved especially when the Gaia EDR3 CMD is analysed since main evolutionary sequences are visible: the brightest MS stars and the MSTO point (at G = 18.8 mag), a compact RC (at K = 13.51 mag and at G = 16.3 mag) and a well-defined RGB. The main parameters that we estimated for this cluster are extinctions of A_Ks = 0.09 mag and A_G = 0.89; a distance modulus of (m − M)₀ = 15.12 mag, corresponding to D = 10.6 kpc, in good agreement with D_GEDR3 = 9.6 kpc from the Gaia passband; and a galactocentric distance of R_G = 11.35 kpc. Fitting the PARSEC isochrone models we found that ESO 92-18 may be an old OC or young GC with an age = 5 ± 1 Gyr and a metallicity of [Fe/H] = − 0.9 ± 0.2 dex, which are in excellent agreement with the parameters derived from the χ² analysis. Our results agree with the parameters obtained by Salaris et al. (2004) and Shao & Li (2019), especially when ages are compared. Bica et al. (2019) also classified ESO 92-18 as an OC.

5.12. Gaia 2

Gaia 2 is inspected matching the Gaia EDR3 + 2MASS datasets. Although the 2MASS CMD is not as clear to interpret, the Gaia EDR3 CMD shows the MSTO point at G = 17.7 mag as a clear overdensity, the highest part of the MS stage, but also a poorly populated RGB (11 stars). We can use these sequences in order to determine its main physical parameters. We estimate an extinction of A_Ks = 0.10 mag and A_G = 0.97 mag, a distance modulus of (m − M)₀ = 13.43 mag, which can be translated into a heliocentric distance of D_NIR = 4.91 kpc, in very good agreement with D_GEDR3 = 4.43 kpc, and a galactocentric distance of R_G = 12.03 kpc. The best isochrone fit yields an age = 10 ± 1 Gyr and a metallicity of [Fe/H] = − 0.9 ± 0.2 dex, which are in good agreement with the values obtained from the χ² analysis. This time the uncertainties on the metallicity are higher than those on the age due to the lack of evolved stars (especially for the RC or BHB), but also because the MSTO point is less sensitive to metallicity changes. Additionally, since this cluster is near the Sun, we calculate its distance using the Gaia EDR3 parallax values. We use the Bayesian distances provided by Bailer-Jones et al. (2021), finding a mean value of D = 3.57 ± 1.08 kpc, in good agreement with the previous estimate, within 1σ.

Moreover, we compare our results with Bica et al. (2019), which catalogued this cluster as an ultra-faint GC (M_V = −2 mag), located at heliocentric distance of 5.4 kpc, in agreement within 1σ with our estimates. Finally, we classify Gaia 2 as a good GC candidate.

6. Summary and discussion

We present the first physical characterisation of 12 newly discovered stellar clusters. A dedicated decontamination procedure was carried out in order to select the PM star members. There are several difficulties usually found in this kind of study, from the differential reddening that distorts the evolutionary sequences in the CMDs (as shown by the FSR 190 CMDs in Fig. B.2) and introduces uncertainties in the observed parameters (e.g., distance estimates) to the high stellar crowding, which may contaminate the final catalogue or alter the total luminosity of the cluster, or to the presence of nebulosity as in the case of Ferrero 54 (Fig. 4). For these reasons our photometric analysis is based on the combination of the optical Gaia EDR3 and NIR VVVX–2MASS datasets in order to obtain reliable parameters and understand the real nature of these low-luminosity objects.

Existing reddening maps in the NIR and in the optical are adopted in order to derive the extinction values. We find a wide range of 0.09 ≲ A_Ks ≲ 0.86 mag in NIR and between 0.89 ≲ A_G ≲ 4.72 mag in the optical. Using these parameters and intrinsic RC colour and magnitude values, we are able to measure the distance modulus for each cluster, translating them into heliocentric distances. We find a large distance range between 4.4 and 18.1 kpc that places them at 2.94 and 14.06 kpc from the Galactic centre.

Ages and metallicities are evaluated adopting the isochrone-fitting method, using the PARSEC model. We find that eight clusters show rich or intermediate metal content with [Fe/H] between −0.5 and −0.9, one young cluster with [Fe/H] = + 0.3, and there are two are metal-poor clusters with [Fe/H] = − 1.4 and −1.8. Regarding the age, for most of our clusters we are only able to provide a lower limit, using the vertical method since the MSTO point is below the detection limit.

However, we can better constrain the ages of the following clusters: Kronberger 100, Kronberger 119, Kronberger 143, Patchick 125, ESO 92-18, and Gaia 2. As shown in Fig. 11, we find clusters with a wide range of ages. The majority of clusters populate the Galactic thin and thick disk and exhibit ages from 2 Gyr to 13 Gyr. However, we also identify two old clusters (Patchick 125 and Patchick 126) that are located well inside the Galactic bulge. Their ages of 14 Gyr and > 8 Gyr, respectively, are in line with the results from previous works, implying that the lower age limit for the MW bulge is ∼5 Gyr (e.g., Clarkson et al. 2011) and the mean age is ∼10 Gyr (e.g., Freeman 2008), suggesting that the clusters are either real tenants of the bulge or, alternatively, intruders originating from the halo (e.g., Forbes & Bridges 2010; Massari et al. 2019; Minniti et al. 2021a; Romero-Colmenares et al. 2021).

Fig. 11. Age as function of NIR heliocentric distance (top panel) and of galactocentric distance RG (bottom panel) for our sample. Also show are the age error bars, even if not all clusters have error bars in their ages in Table 2. However, the lower limit given in Table 2 is adopted as the lower error, and the higher limit of 13 Gyr, the typical age of old Galactic GCs, is adopted as the higher error. The age error is thus the difference between our age estimates and 13 Gyr.

In conclusion, we confirm the OC nature for Kronberger 100, while we classify Kronberger 119 and Kronberger 143 as old OCs based on their comparably young ages and their relative co-movement with the surrounding stellar field, although a classification as young GCs cannot be ruled out. For the same reasons ESO 92-18, the most populated cluster analysed, is also categorised as an old OC or young GC. Five objects are classified as GCs based on our findings: Patchick 125 and Riddle 15 are classified as a metal-poor GCs, Patchick 126 and Gaia 2 as GCs of intermediate metallicity, and Ferrero 54 as metal-rich GC. Finally, we cautiously categorise three objects with poorly constrained ages (Kronberger 99, Patchick 122, and FSR 190) as GC candidates.

We note that in all cases deeper Hubble Space Telescope (HST) or James Webb Space Telescope (JWST) observations as well as spectroscopic observations are needed in order to build complete catalogues of likely cluster members, improving the structural parameters and luminosities.

Acknowledgments

We gratefully acknowledge the use of data from the ESO Public Survey program IDs 179.B-2002 and 198.B-2004 taken with the VISTA telescope and data products from the Cambridge Astronomical Survey Unit. ERG acknowledges support from ANID PhD scholarship No. 21210330. D.M. gratefully acknowledges support by the ANID BASAL project FB210003. 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). R.K.S. acknowledges support from CNPq/Brazil through project 305902/2019-9. J.A.-G. acknowledges support from Fondecyt Regular 1201490 and from ANID – Millennium Science Initiative Program – ICN12_009 awarded to the Millennium Institute of Astrophysics MAS.

References

Appendix A: Optical images for the star clusters

Fig. A.1. Optical images of the sky regions centred on the star clusters analysed in the present work: (a) Kronberger 99, (b) Kronberger 100, (c) Kronberger 119, (d) Kronberger 143, (e) Patchick 122, (f) Patchick 125, (g) Patchick 126, (h) ESO 92-18, (i) Ferrero 54, (l) FSR 190, (m) Gaia 2, and (n) Riddle 15. DECaPS colours were used for most of the targets, except for Riddle 15 and FSR 190 for which PanSTARRS colours were used. A coverage of 5.50′×6.25′ was used for the most of the clusters, except for FSR 190 (11.5′×13′).

Appendix B: VPMs and CMDs for the cluster sample

Fig. B.1. VPM diagrams (Table 1). The brightest cloud represents the highest PM probability (within a radius < 1 mas yr−1), which indicates the presence of the cluster, as described in Sect. 4.1.

Fig. B.2. CMDs for the PM member stars. Left and middle panels: Optical Gaia EDR3 CMDs. A family of PARSEC isochrones was fit in order to derive the uncertainties on the age and on the metallicity estimates. Right panel: NIR (VVVX and/or 2MASS) CMDs. In all panels the red lines depict the PARSEC isochrones with age and metallicity listed in Table 2 for each cluster. In the CMDs the cyan areas represent overdensities, whereas the black areas are underdensities.

Fig. B.2. (continued)

Fig. B.2. (continued)

All Tables

All Figures

	Fig. 1. Galactic distribution of GCs towards the MW bulge. Star clusters investigated in this work are highlighted with red points, while open green squares represent the Galactic GCs by Vasiliev & Baumgardt (2021) in the Gaia EDR3 footprint.
In the text

	Fig. 2. Gaia EDR3 density maps, cleaned from nearby stars, centred on the ESO 92-18 (left) and Kronberger 100 (right) fields. These two clusters are used to visually show the different target sizes. The greener areas are higher densities, while the bluer areas are lower densities.
In the text

Fig. 3. KDE technique applied to select the likely size of a cluster during the decontamination procedure. From left to right: scatterplot, 2D histogram, Gaussian KDE, and 2D density with contours (or iso-densities) are shown for the ESO 92-18 (top panels) and Kronberger 100 (bottom panels) fields, used as examples. The green and yellow areas are indicate higher densities, while the blue areas show lower densities.

In the text

	Fig. 4. Optical DECaPS (left) and NIR VVVX (right) images of the sky region centred on Ferrero 54. The field of view is 200″ × 200″, oriented along the Galactic coordinates, with Galactic longitude increasing to the left and Galactic latitude increasing to the top. Veils of gas and dust are evident as green filaments in the optical passband. They are part of the Vela supernova remnant complex.
In the text

	Fig. 5. Patchick 125 (bottom panels) and Patchick 126 (top panels) NIR CMDs. In the left panels we compare two isochrones using two different extinction coefficients: RV = 3.1 (yellow points) and RV = 2.5 (red points). In the right panels we fix the RV = 2.5 and the ages as listed in Table 2, and we change the metallicity as specified in the legend.
In the text

	Fig. 6. Galactocentric distribution of star clusters analysed in the present work (red points) and Galactic GCs (green open squares) by Baumgardt & Vasiliev (2021). The black cross represents the position of the Galactic centre, while the blue star is the position of the Sun.
In the text

	Fig. 7. Gaia EDR3 PM distribution of star clusters in the present work (red points) and Galactic GCs (open green squares) by Vasiliev & Baumgardt (2021).
In the text

Fig. 8. χ2 estimate as a function of age for three representative clusters: ESO 92-18, Patchick 125, and Patchick 126. The colour of the lines depends on the metallicity, specified in the legend. The black dashed lines indicate the minimum χ2 value at the relative age. The age and metallicity that correspond to the ${\chi }_{min}^2$ are listed in Table 3. Right panel: Gaia EDR3 CMD for ESO 92-18. The PARSEC isochrone (black dotted line) was fit using the age and metallicity obtained by the χ2 analysis and listed in Table 3. The reddening, distance modulus, and extinction are listed in Table 2.

In the text

	Fig. 9. Gaia EDR3 optical CMDs for the clusters: Kronberger 99 (left panel), Kronberger 143 (middle panel), and ESO 92-18 (right panel). The green, yellow, and pink lines are the Dartmouth isochrones with [α/Fe]= + 0.4, +0.2, 0, respectively. The dotted red lines are the PARSEC isochrones. We keep the age and metallicity fixed as listed in Table 2.
In the text

Fig. 10. Position (top left panel), magnitude-separation (top right panel), Gaia EDR3 PM-distribution (bottom left panel), and period-amplitude (bottom right panel) diagrams for the RR Lyrae stars located within 20′ from the centre of the cluster Patchick 125. The mean ridge lines for the Oosterhoff type I and II populations (left and right dotted lines, respectively) are from Clement & Shelton (1999). The two RR Lyrae members are shown as blue points.

In the text

Fig. 11. Age as function of NIR heliocentric distance (top panel) and of galactocentric distance RG (bottom panel) for our sample. Also show are the age error bars, even if not all clusters have error bars in their ages in Table 2. However, the lower limit given in Table 2 is adopted as the lower error, and the higher limit of 13 Gyr, the typical age of old Galactic GCs, is adopted as the higher error. The age error is thus the difference between our age estimates and 13 Gyr.

In the text

Fig. A.1. Optical images of the sky regions centred on the star clusters analysed in the present work: (a) Kronberger 99, (b) Kronberger 100, (c) Kronberger 119, (d) Kronberger 143, (e) Patchick 122, (f) Patchick 125, (g) Patchick 126, (h) ESO 92-18, (i) Ferrero 54, (l) FSR 190, (m) Gaia 2, and (n) Riddle 15. DECaPS colours were used for most of the targets, except for Riddle 15 and FSR 190 for which PanSTARRS colours were used. A coverage of 5.50′×6.25′ was used for the most of the clusters, except for FSR 190 (11.5′×13′).

In the text

	Fig. B.1. VPM diagrams (Table 1). The brightest cloud represents the highest PM probability (within a radius < 1 mas yr−1), which indicates the presence of the cluster, as described in Sect. 4.1.
In the text

Fig. B.2. CMDs for the PM member stars. Left and middle panels: Optical Gaia EDR3 CMDs. A family of PARSEC isochrones was fit in order to derive the uncertainties on the age and on the metallicity estimates. Right panel: NIR (VVVX and/or 2MASS) CMDs. In all panels the red lines depict the PARSEC isochrones with age and metallicity listed in Table 2 for each cluster. In the CMDs the cyan areas represent overdensities, whereas the black areas are underdensities.

In the text

	Fig. B.2. (continued)
In the text

	Fig. B.2. (continued)
In the text