Fundamental parameters for 45 open clusters with Gaia DR2, an improved extinction correction and a metallicity gradient prior
H. Monteiro, W. S. Dias, A. Moitinho, T. Cantat-Gaudin, J. R. D. Lépine, Carraro, E. Paunzen
MMNRAS , 1–19 (2020) Preprint 25 September 2020 Compiled using MNRAS L A TEX style file v3.0
Fundamental parameters for 45 open clusters with GaiaDR2, an improved extinction correction and a metallicitygradient prior
H. Monteiro, (cid:63) , W. S. Dias , A. Moitinho , T. Cantat-Gaudin, , J. R. D. L´epine ,G., Carraro, and E. Paunzen Instituto de F´ısica e Qu´ımica, Universidade Federal de Itajub´a, Av. BPS 1303 Pinheirinho, 37500-903 Itajub´a, MG, Brazil CENTRA, Faculdade de Ciˆencias, Universidade de Lisboa, Ed. C8, Campo Grande, 1749-016 Lisboa, Portugal Institut de Ciencies del Cosmos, Universitat de Barcelona (IEEC-UB), Marti i Franques 1, E-08028 Barcelona, Spain Universidade de S˜ao Paulo, Instituto de Astronomia, Geof´ısica e Ciˆencias Atmosf´ericas, S˜ao Paulo, SP, Brazil Department of Physics and Astronomy, University of Padova, Vicolo dell’Osservatorio 3, I-35122 Padova, Italy 0000-0002-0155-9434 Departament of Theoretical Physics and Astrophysics. Marsaryk University. Brno, Czech Republic
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
Reliable fundamental parameters of open clusters such as distance, age and ex-tinction are key to our understanding of Galactic structure and stellar evolution. Inthis work we use
Gaia
DR2 to investigate 45 open clusters listed in the
New catalogueof optically visible open clusters and candidates (DAML) but with no previous astro-metric membership estimation based on
Gaia
DR2. In the process of selecting targetsfor this study we found that some clusters reported as new discoveries in recent papersbased on
Gaia
DR2 were already known clusters listed in DAML. Cluster membershipswere determined using a maximum likelihood method applied to
Gaia
DR2 astrome-try. This has allowed us to estimate mean proper motions and mean parallaxes for allinvestigated clusters. Mean radial velocities were also determined for 12 clusters, 7 ofwhich had no previous published values. We have improved our isochrone fitting codeto account for interstellar extinction using an updated extinction polynomial for the
Gaia
DR2 photometric band-passes and the Galactic abundance gradient as a prior formetallicity. The updated procedure was validated with a sample of clusters with highquality [ Fe / H ] determinations. We then did a critical review of the literature and ver-ified that our cluster parameter determinations represent a substantial improvementover previous values. Key words: (Galaxy:) open clusters and associations:general
The fundamental parameters of open clusters (OCs) - dis-tance, age, metallicity, interstellar extintion along the lineof sight, proper motions and radial velocities - have longbeen considered key for revealing the structure and evolutionof the Milky Way (Becker & Fenkart 1970; Janes & Adler1982). However, because each cluster contributes with a sin-gle point in parameter space, the accumulation of OC datahas traditionally been a lengthy process, with leaps in ourknowledge of the Galaxy based on OCs taking many years(Moitinho 2010). Such a jump has been recently broughtby the ESA Gaia mission (Gaia Collaboration et al. 2016). (cid:63)
E-mail:[email protected]
The
Gaia
Data Release 2 catalogue (DR2, Gaia Collabo-ration et al. 2018a) provides precise astrometric and pho-tometric data for more than one billion stars with magni-tude G brighter than 21, which are bringing a new era ofGalactic research with OCs. A summary of various past,pre- Gaia , efforts to compile homogeneous OC parameters isgiven in (Netopil et al. 2015) and a review of pre-
Gaia resultsof Galactic structure with OCs can be found in Moitinho(2010).The richness of
Gaia
DR2 has triggered numerous largescale OC studies. Without being exhaustive, we indicatesome significant examples: Cantat-Gaudin et al. (2018) andCantat-Gaudin & Anders (2020) determined proper motionsand distances for 1481 open clusters based on membershipobtained using the UPMASK membership determination © a r X i v : . [ a s t r o - ph . S R ] S e p H. Monteiro et al. method (Krone-Martins & Moitinho 2014). Soubiran et al.(2018) determined proper motions and radial velocities fora kinematic study of 406 OCs. Liu & Pang (2019) used theFriend of Friend method to flag over two thousand clustercandidates. Kounkel & Covey (2019) performed a cluster-ing analysis to study 1900 possible aggregates within 1 kpc.Also in the solar neighborhood, Sim et al. (2019) reportedon 655 clusters (proposing 207 new candidates) by visual in-spection of the stellar distributions in proper motion spaceand spatial distributions in the Galactic coordinates ( l , b ) space. Members were determined using Gaussian mixturemodel and mean-shift algorithms. Monteiro & Dias (2019)determined the parameters of 150 OCs adopting a maxi-mum likelihood method to estimate cluster memberships.Using the same procedure Dias et al. (2019) determined theparameters of several hundreds of OCs, from which they se-lected 80 younger that 50 Myr for determining the spiral pat-tern rotation speed of the Galaxy and the corotation radius.Bossini et al. (2019) employed a Bayesian methodology fordetermining the ages, distances and interstellar absorptionfor 269 OCs with membership determinations from Cantat-Gaudin et al. (2018). Castro-Ginard et al. (2020), using adeep learning artificial neural network (ANN), reported thediscovery of 588 new OCs for which they estimated distancesand proper motions. Likewise using an ANN to characterise1867 OCs, Cantat-Gaudin et al. (2020) analysed the spiralstructure, scale height of the thin disk and warp of the MilkyWay. It is also worthwhile mentioning that Gaia
DR2 hasalso been used in combination with ground based observa-tions for smaller scale, but more detailed studies of individ-ual objects (e.g. Dias et al. 2018; Perren et al. 2020).Despite the intense activity enabled by the high qual-ity
Gaia
DR2 data, many previously known objects remainwith no membership and parameter determinations basedon Gaia DR2. The goal of this paper is to present our deter-minations of the fundamental parameters of these difficultleft-over clusters and the methodological improvements thatallowed to reach those results.The remainder of the manuscript is organized as follows.In the next section, we describe the data selection and thesample of the studied objects. Section 3 is dedicated to de-scribe the method of astrometric membership determinationand to briefly introduce the isochrone fitting procedure. Insection 4 we present improvements to our isochrone fittingprocedure using a revised treatment of interstellar extinctionwith updated
Gaia photometric band-passes and constrain-ing metallicity. These improvements are validated with acontrol sample of clusters from the literature. In section 5we discuss the results and in section 6 we compare the val-ues here obtained with those from the literature. Finally, insection 7 we give some concluding remarks.
We started by cross-matching all 2167 clusters published inthe
New catalog of optically visible open clusters and candi-dates (Dias et al. 2002, hereafter DAML) with the literaturefor which membership determinations using
Gaia
DR2 datawere available (Cantat-Gaudin et al. 2018; Castro-Ginardet al. 2020; Liu & Pang 2019; Sim et al. 2019). This led toa list of 75 clusters for which no previous
Gaia
DR2 based memberships were available. For each cluster we selected thestars in
Gaia
DR2, using the central coordinates and the ra-dius taken from the DAML catalogue. To allow for someuncertainty in the radius and include possible cluster mem-bers further away from center, we took a region in the skywith radius 2 arcmin larger than the radius listed in DAML.We note that stars originated in the cluster might be fur-ther away due to processes such as dynamical evolution oran underestimated radius. However, for the purposes of thiswork, complete samples of members are not required, butonly enough stars for determining the reddening, distanceand age of the clusters. Before determining the astrometricmembership as detailed in the next section, we filtered thedata to assure that only reliable astrometric solutions wereused. The filtering was done following the recipe publishedby Gaia Collaboration et al. (2018b), which takes into ac-count systematic effects of
Gaia data, consistency between G and G BP + G RP filter fluxes, as well as the number ofpasses in the given field, among other factors. As describedin section 5, a subsequent quality control step left us with afinal sample of 45 clusters for which results are presented. The membership analysis follows the method described inDias et al. (2014). We assume that errors in proper motioncomponents and parallaxes are normally distributed and usea maximum likelihood method to obtain the membershipsadopting a model which assumes Gaussian distributions forproper motions in both cluster and field stars. The model isdescribed by equation 1 where the uncertainties of the dataand their correlations follow the recommendation given byLuri et al. (2018) such that the probability f ( X ) is given by: f ( X ) = exp (cid:16) − ( X − µ ) T Σ − ( X − µ ) (cid:17)(cid:112) ( π ) k | Σ | (1)where X is the column vector ( µ α cos δ , µ δ , (cid:36) ) composedof the proper motion components and the parallax, µ themean column vector and | Σ | is the co-variance matrix, whichincorporates the uncertainties ( σ ) and their correlations ( ρ ),given in the Gaia
DR2 catalogue.The maximum likelihood solution provides the distri-bution of cluster membership probabilities. This allows thedetermination of the cluster membership probability of eachstar in the selected field as well the mean proper motionsand parallaxes of the clusters, considering as members thosestars with cluster membership probability greater than 0.51.The adopted membership cut-off of 0.51 is merely based onthe availability of statistical evidence for the pertinence to agiven cluster and used as a compromise between complete-ness and contamination. As discussed in the next section, theisochrone fitting procedure will use the membership proba-bilities for decreasing the weight of the possible contami-nants in the determination of the cluster fundamental pa-rameters. Still, For the open clusters studied here we alsoran the fits with a cut-off of 0.8 as a sanity check on the re-sults. The differences with respect the results obtained withthe 0.51 cut-off were ( . ± . ) dex, ( − . ± . ) pc, MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 ( − . ± . ) mag, for age, distance and A V respectively,which are are comparable to the uncertainties obtained ineither case, showing that adopting one or the other cut-offis equivalent within the errors.We also estimate radial velocity as the mean of the ra-dial velocity data with a σ outlier rejection of the members.We note that Gaia DR2 radial velocities are only availablefor small numbers of cluster members. The estimated un-certainty is given by the standard deviation of the radialvelocities. It is well known that the stars in an open cluster align alonga distinctive sequence in a color-magnitude diagram (CMD).This sequence is most evident when only stars with a suf-ficiently high stellar membership probability (e.g. as deter-mined by the method described above) are included. In otherwords, the sequence is most evident when field star contam-ination is minimum. Likewise, the member stars that formthis feature should exhibit a clump in a 3D plot with propermotion and parallax data, since they occupy a limited vol-ume in space and have similar velocities. In this context, anet evidence of a cluster sequence in a CMD of member starsis a strong indicator of the presence of a real open clusterand allows the determination of its age, extinction, and anestimate of the cluster distance independent of the parallaxmeasurements. Consequently, the next step in our analysiswas to use
Gaia
DR2 G BP and G RP magnitudes and to per-form the isochrone fits to the cluster member stars identifiedwith the method outlined above.As discussed in previous works (e.g. Dias et al.2018), membership knowledge and an objective method forisochrone fitting are determinant to the final results. Wenote that many isochrone fits performed in the literature,objective or not, were based on limited membership deter-minations, mainly due to large errors or even absence ofstellar proper motions and/or parallax data.Here we applied the cross-entropy (CE) method to fittheoretical isochrones to the CMDs of cluster member starsas detailed in Monteiro et al. (2017). This approach has al-ready been successfully applied to Gaia
DR2 data in Diaset al. (2018), Monteiro & Dias (2019) and Dias et al. (2019).In short, the CE method involves an iterative statistical pro-cedure where in each iteration the initial sample of the fitparameters is randomly generated using predefined criteria.Then, the code selects the 10 % best fits based on calcu-lated weighted likelihood values taking into account the as-trometric membership probabilities. Based on the parameterspace defined by the best fits, a new random fit parametersample is generated and applied in the following run of thecode. This procedure continues until a convergence criterionis reached. In other words, the isochrone fit in this techniqueconsists in choosing the best set of points of a model withrespect to the set of points of the observed data. The errorsof the fit are estimated by bootstrapping the process. Thisalso reduces the influence of possible field stars contaminat-ing the lists of members. In our code we adopt a likelihood function given in theusual manner for the maximum likelihood problem as: L( D N | X ) = N (cid:214) i = Φ ( I ( X ) , D N ) , (2)where Φ ( I ( X )) is a multivariate normal, X is the vectorof parameters ( A V , distance d , age lo g ( t ) and [ Fe / H ] ), I ( X ) is the synthetic cluster obtained for the isochrone defined by X and D N the data for the N observed stars in the cluster.The likelihood function above is used to define the ob-jective function as: S ( X | D N ) = − lo g ( P ( X ) × L( D N | X )) (3)where the function P ( X ) is the prior probability forthe parameters given by P ( X ) = (cid:206) nn = P ( X n ) . For age weadopt P ( X n ) = , for distance we use N( µ, σ ) obtained withBayesian inference from the parallax ( (cid:36) ) and its uncertainty( σ (cid:36) ) and the variance ( σ ) is obtained from the distanceinterval calculated from the inference using the uncertaintyas σ (cid:36) . The prior in A V is also adopted as a normal dis-tribution with µ and variance ( σ ) for each cluster takenfrom the 3D extinction map produced by Capitanio et al.(2017) . The prior for [ Fe / H ] used the the Galactic gradi-ent from Donor et al. (2020) as detailed in the Section 4.2.The optimization algorithm then minimizes with respect to X . In the present study, our algorithm uses the PadovaPARSEC version 1.2S database of stellar evolutionary tracksand isochrones (Bressan et al. 2012), which uses the Gaia fil-ter passbands of Ma´ız Apell´aniz & Weiler (2018), is scaled tosolar metal content with Z (cid:12) = . and scans the followingparameter space limits: • age: from log(age) =6.60 to log(age) =10.15; • distance: from 1 to 25000 pc; • A V : from 0.0 to 5.0 mag; • [ Fe / H ] : from -0.90 to +0.70 dexSince our method uses a synthetic cluster obtained frommodel isochrones, we include the extinction for each stargenerated based on a A λ / A V relation of choice. For eachgenerated star of the synthetic cluster we obtain, in eachfilter, what would be the reddened observed photometry forthe particular model I ( X ) . The synthetic clusters have beengenerated with a binary fraction of 0.5 and the masses ofcomponents drawn from the same IMF. The synthetic clus-ter is then compared to the observed data through the like-lihood defined in Eq. 2. When analyzing the clusters with the software of Monteiroet al. (2017) described in the previous section, we noticedthat about 20% (8 clusters) of the fits would only convergeto consistent solutions when only G BP and G RP magnitudes The 3D extinction map is available online at https://stilism.obspm.fr/
MNRAS000
MNRAS000 , 1–19 (2020)
H. Monteiro et al. were used, without using G . For most of these clusters theextinction was considerable, reaching as high as A V = . .We had originally adopted the same polynomial as Bossiniet al. (2019) to correct for extinction, although they only in-vestigated clusters with low A V and used the now outdatedband passes. Therefore, we decided to redo the extinctionpolynomial based on the updated Gaia filter band-passes.In the process, we analyzed different approaches for con-straining another key parameter: metallicity.
To account for the extinction coefficients dependency oncolour and extinction due to the large passbands of
Gaia fil-ters, we followed the procedure described by Danielski et al.(2018) and used in Gaia Collaboration et al. (2018b). Weused the same model atmospheres and same value grid: Ku-rucz model spectra (Castelli & Kurucz 2003) (for 3500 K < T eff < g =2.5 and 4. For the extinction law we adopted themore recent one from Fitzpatrick et al. (2019) and a gridof 0.01 < A V < Gaia photometric passbands, giventhat these bands provide better agreement between syn-thetic
Gaia photometry and
Gaia observations.The model spectra were convolved with the filter pass-bands and extinction scaled law to construct a grid of red-dened photometry. The extinction coefficients k m were cal-culated with the equations below: A m = m − m = − . (cid:32) ∫ F λ T λ E A V λ d λ ∫ F λ T λ d λ (cid:33) (4)and k m = A m A V (5)where E A V λ is the extinction function, which in this casewas the Fitzpatrick et al. (2019) law.A polynomial defined as in Eq. 6 was then fit to the k m versus A V grid of values using the package (The AstropyCollaboration et al. 2018). In this expression x and y are A V and G BP − G RP , respectively, k m is the extinction coefficientand the m subscript refers to each of the bands G , G BP and G RP . Unlike the work in Gaia Collaboration et al. (2018b),here we fit a full 4th degree polynomial to the grid. Theresults of the fit are given in Table 1. k ( x , y ) = c + c x + ... + c n x n + c y + ... + c n y n + c x y + c x y + ... + c ( n − ) x y n − + ... + c ( n − ) x n − y (6)Our results agree with the ones obtained by Wang &Chen (2019), using a different method. Specifically, they de-rive their own extinction law and do not fit a polynomial tothe A V -color dependence, but do apply corrections for thelarge filter passbands. They obtain . ± . for k BP and . ± . for k RP . Our average results from the polyno-mial fit are . ± . and . ± . for k BP and k RP respectively. For the G filter we get . ± . while Wang& Chen (2019) obtained . ± . for k G . To validate and to determine possible limitations of the newextinction polynomial we have applied our code to a sampleof well studied clusters. The sample was defined with clus-ters that had [ Fe / H ] determined from high resolution spec-troscopy in Netopil et al. (2016) as well as from APOGEEas published in Carrera et al. (2019). Both samples have agood coverage of the fundamental parameters age, distanceand A V .We performed four test runs of our fitting procedure: 1)using a prior on distance and A V only; 2) using a prior indistance, A V and [ Fe / H ] based on the Galactic abundancegradient from Donor et al. (2020); 3) using a prior in dis-tance, A V and [ Fe / H ] fixed at values from Carrera et al.(2019) and 4) using a prior in distance, A V and [ Fe / H ] fixedat the solar value.A first consistency check is to see how the fundamen-tal parameters age, distance and A V are affected by fixingor not the parameter [ Fe / H ] . This is important for assess-ing the degree to which the fit results are sensitive to theassumptions made. In Fig. 1 we show the comparison of re-sults obtained for the fundamental parameters with [ Fe / H ] held fixed at the value from Carrera et al. (2019) and al-lowed to vary subjected to the Galactic metallicity gradientprior. The results show that the agreement is good betweenparameters determined using both strategies. There are nodetectable systematic effects. Considering the fact that A V and [ Fe / H ] are generally hard to untangle based on pho-tometry, this is an indication that the high quality of Gaia photometry allows for a good definition of CMD shape andthis removes some of the degeneracy in these parameters.Then we look at how the discrepancies in parameter es-timates obtained from fits using a prior for [ Fe / H ] based onthe Galactic abundance gradient and fits using [ Fe / H ] = . (which is the usual procedure adopted when this parameteris unknown), when compared to estimates obtained fromfits where [ Fe / H ] is fixed to the values from Carrera et al.(2019) which we take to be the most accurate. In Fig. 2 weshow histograms of the discrepancies for log(age), distanceand A V in both situations. The histograms show that as-suming [ Fe / H ] = . leads to slightly larger discrepancies inlog(age) and similar in distance although some outliers areclearly seen. These outliers are all from clusters with CMDsor turn-offs that are not clearly defined. There is a smallsystematic overestimation of 0.05 mag in A V as well.The sensitivity of Gaia data to [ Fe / H ] can be verifiedin the results shown in Fig. 3 where the metallicity valuesobtained from fits using the Galactic abundance gradientprior are compared to values from Carrera et al. (2019) andNetopil et al. (2016). The same behaviour was found for fitswhere [ Fe / H ] had no prior albeit with a larger spread, asexpected. The average differences from the literature valuesare . ± . and . ± . with respect to Carreraet al. (2019) and Netopil et al. (2016) respectively. Based onthis result we incorporate a baseline error of 0.15 which iscombined quadratically with the fit error to give the finaluncertainty in [ Fe / H ] .It is important to point out that this [ Fe / H ] estimate MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 Table 1.
Coefficients of the polynomial fit to the k m versus A V grid of values. Band c c c c c c c c c c c c c c c G BP G RP G Figure 1.
Comparison of results obtained for the fundamental parameters with [ Fe / H ] held fixed at the value from Carrera et al. (2019)and allowed to vary with a prior based on the Galactic metallicity gradient as described in the text. Figure 2.
Discrepancies in parameter estimates, obtained from fits using a prior for [ Fe / H ] based on the Galactic metallicity gradientand fits using [ Fe / H ] = . when compared to estimates obtained from fits where [ Fe / H ] is fixed to the values from Carrera et al. (2019). should not be used as a proper metallicity determination forthe open clusters studied. While the derived values of [ Fe / H ] are indicative of the metallicity of individual clusters, statis-tically they are based on the metallicity gradient prior andthus cannot be used as a set for determining the Galacticabundance gradient. We have chosen to use [ Fe / H ] as a freeparameter because, as discussed above, it gives less biasedresults for A V when compared to the widespread practiceof adopting [ Fe / H ] = . . By letting [ Fe / H ] vary as a freeparameter we also get more reliable estimates and uncer-tainties in the other parameters. Another positive point inadopting this strategy is that it may indicate clusters whereinteresting or deviant properties may be present allowing fora sample selection for more detailed observational followupcampaigns.As shown above, compared to the fixed [ Fe / H ] priorfrom high resolution spectroscopy, adopting the [ Fe / H ] val-ues determined with the abundance gradient prior does notintroduce systematic effects in the other parameters. Basedon these results we have adopted the following procedure forthe fits in this work: 1) if there is a reliable determination of [ Fe / H ] in the literature, such as in Carrera et al. (2019) andNetopil et al. (2016) we adopt that value and its uncertaintyfor the metallicity prior; 2) if there are no reliable value tobe used as prior we use a prior based on the Galactic metal-licity gradient from Donor et al. (2020). The results of theisochrone fits, using the Galactic metallicity gradient prior,to the clusters with high resolution spectroscopy analyzedin this section are given in Table A1. Of the 75 clusters selected as described in section 2, themembership results for 30 objects either did not reveal iden-tifiable cluster sequences or the isochrone fits were not agood match to the data and were thus discarded from fur-ther consideration. These clusters are identified in Table B1.Typically, the fits failed when the sequences were faint andtherefore had a small magnitude range, with higher errors,for fitting. While it may seem that our method is not beingable to produce results for a high fraction of the clusters, we
MNRAS000
MNRAS000 , 1–19 (2020)
H. Monteiro et al.
Figure 3.
Comparison of [ Fe / H ] estimates obtained from isochrone fits using the Galactic metallicity gradient as prior to values obtainedfrom the literature. Left panel shows comparison to values from APOGEE from Carrera et al. (2019), where symbols are colored accordingto A V . No systematic deviations due to A V are apparent. Right panel shows comparison to values from Netopil et al. (2016) indicatingerrors as described in the text and discriminating between the high quality sample (HQS) and the lower quality sample (LQS) as definedby the authors. Figure 4.
Example of a rediscovered star cluster. Field of (cid:48) × (cid:48) centered on Hogg 22, with coordinates from Dias et al. (2002).The members of the cluster labelled UBC 547 by Castro-Ginardet al. (2020) are marked in red. In the upper right (N-W) we startseeing a concentration of bright stars which are on the edge of thenearby open cluster NGC 6204. note that the selected sample of 75 clusters is composed ofthe leftovers from previous works. Thus, in fact our pipelinehas been able to successfully handle 45 remaining consider-ably difficult cases. The classifications from DAML illustratethe type of objects in our sample: 10 were discovered in infra-red but are visible in the DSS images; Dolidze 1 was classifiedas possible cluster; Dolidze 35 and ESO 332-13 as a dubiousobjects; ESO 392-13 was an not found in the DSS imagesinspection; Sigma Orionis, NGC 1977, and Trapezium were classified as possible OB associations, and ESO 429-02 wasclassified as a possible open cluster remnant.With respect to the Trapezium cluster, the situation isfurther complicated by the presence of other young stellarpopulations along the same line of sight (Alves & Bouy 2012;Chen et al. 2019). The cluster here studied is composed ofoptically revealed, low extinction, elements in the foregroundof the embedded Trapezium cluster. It is part of foreground”group 5” in Chen et al. (2019), which includes NGC 1980and NGC 1981. In this work we identify this stellar aggregateas Trapezium-FG.During the analysis we found that some clusters re-ported as new discoveries in recent papers were known clus-ters listed in DAML. The clusters FoF 2316 and FoF 868found by Liu & Pang (2019) using using Gaia
DR2 havesimilar positions, parallaxes and proper motions, and coin-cide with NGC 6530. We note, however, that they are notin the high quality ”Class 1” group defined by those au-thors. Castro-Ginard et al. (2020), also using
Gaia
DR2, re-ported the discovery of 582 clusters which they identify un-der the designation ”UBC”. Some are known clusters listedin DAML: Czernik 43 = UBC 399; Dolidze 1 = UBC 367;ESO 429 02 = UBC 464; FSR 0761 = UBC 197; Hogg 22 = UBC 547; IC 1442 = UBC 164; NGC 133 = UBC 185;NGC 1977 = UBC 621; NGC 1980 = UBC 208; NGC 6444 = UBC 329; Ruprecht 118 = UBC 313. Fig. 4 illustrates thecase of Hogg 22 (UBC 547). Curiously, although Castro-Ginard et al. (2020) mention that NGC1980 is listed inDAML, it is included in the list of newly discovered clus-ters as UBC 208. We note however that these cases arein small number and do not raise concerns over the muchbroader scope of findings in Castro-Ginard et al. (2020).They do however highlight how delicate it is to cross-identifyopen clusters, which are extended objects, not continuouslike galaxies, but often sparse discrete groups with irregularshapes, different apparent sizes and without clear bound-aries.For Berkeley 64 we estimated better central coordi-
MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 nates at α = h m s ; δ = + o (cid:48) (cid:48)(cid:48) in J2000. ForIC 1442 improved central coordinates are α = h m s ; δ = + o (cid:48) (cid:48)(cid:48) in J2000, similar to the value estimated by(Maurya et al. 2020).In the final analysis we also visually checked the color-magnitude diagrams with the isochrone fitted to the G BP and G RP photometric data from Gaia
DR2 catalogue. Thevector proper motion diagram constructed with individualsymbol sizes and colours scaled to the kernel density esti-mated density in proper motion and parallax space as dis-cussed in Monteiro & Dias (2019), was also checked since in( µ α cos δ , µ δ , (cid:36) ) space a real clusters must show a concen-tration of stars.In Table 2 we present the mean astrometric parameters( µ α cos δ , µ δ , (cid:36) ) provided by the method described in section3.1. In Table 3 the parameters obtained by the isochrone fitare given. In Fig. 7 the final results of the isochrone fit withthe stars with membership probability greater than 0.51 areshown.We point out that the fitting procedure has limita-tions in the treatment of very young clusters: On the onehand, variable extinction and age spread within the clus-ter are not specifically included in the fitting model. On theother hand, the grid of PARSEC isochrones does not includeages younger than ∼ log (age) = 6.60) and may notbe particularly suited for pre-main sequence (PMS) evolu-tionary phases. To assess the adequacy of the fits for theyoungest objects, we consider the 12 clusters determinedto be younger than 10 Myr. For 10 clusters (ESO 332 08,ESO 332 13, FSR 0224, NGC 1980, NGC 6530, NGC 6604,Sigma Orionis, Teutsch 132, Trapezium-FG, vdBergh 130),Fig. 7 shows that while the PMS displays some dispersion,this is not evident on the main sequence which has a largefraction of members. The good definition of the main se-quence is indicative that there is no significant variable red-dening or age spread ( (cid:46) ∼
10 Myr should be visually checked andconfirmed. The two remaining clusters in the group (Bica 2,FSR 0236) present a clear dispersion on both the main se-quence and PMS branches. While it is not clear if the dis-persion is due to variable reddening, age spread or a con-tamination of field stars, it is clear that the results for thesetwo clusters may not be very reliable.In Fig. 5 we present the comparison of the distancesobtained with the isochrone fits with those obtained by us-ing the parallax of the member stars. The distances ob-tained from parallaxes were determined with a maximumlikelihood estimation assuming a normal distribution for in-dividual stars and taking into account individual parallaxuncertainties. The standard errors provided in the distance
Figure 5.
Comparison of distances obtained from parallaxes andisochrone fitting, both based on
Gaia
DR2. from parallaxes were estimated by considering a symmetricdistribution so that σ = r − r /( × . ) , which is equiv-alent to 1 σ Gaussian uncertainty, where r and r are 5thand 95th percentile confidence intervals.The comparison shows a good agreement between theparallax distance and the one obtained via isochrone fitting.After 2.5 kpc a slight tendency for larger distances from par-allaxes can be seen, but still within the errors. The resultis a clear improvement with respect to the one presented inMonteiro & Dias (2019). While overall the methods are sim-ilar in both works, the main difference is that here we use arevised Gaia extinction correction and constrain metallicitywith the Galactic abundance gradient prior.The mean difference in the values is of about 218 pcin the sense of distance from parallaxes minus distance fromisochrone fit with a standard deviation of 212 pc. In general,the most discrepant cases are clusters more distant than 5kpc and whose main sequence is defined below G =
16. Inthis region the error in parallax increases considerably goingfrom 0.02 mas ( G ≤
14) to typically 0.15 mas at G = The comparison presented here has two goals: to provide anextra sanity check of our results and to assess the improve-ment they bring. To this end we base the analysis on thewidely used DAML and Kharchenko et al. (2013) (hereafterMWSC) catalogues. It is important to note that these aredifferent types of catalogues. On the one hand, the MWSCis the output of a program applied to the PPMXL (Roeseret al. 2010) and 2MASS (Skrutskie et al. 2006) data. Onthe other hand, DAML is a compilation, curated by hu-mans, of the best results (judged by the curators) availablein the literature. The MWSC aims to overcome the non-uniformity in compilations from the literature, which arebased on results obtained by different authors using differenttechniques, models and calibrations. However, as pointed outin Moitinho (2010) homogeneous methods do not necessarilyproduce the best results. As an example, for close objects,
MNRAS000
MNRAS000 , 1–19 (2020)
H. Monteiro et al.
Figure 6.
Comparison of the values of distance (left panel), age (middle panel) and A V (right panel) obtained by the isochrone fit withthose published in DAML (top) and MWSC (bottom). parallaxes provide the best distances, but at larger distancesisochrone fits are better. Assuming the algorithm employedin the MWSC is flawless, the relatively shallow data usedin the MWSC limits its usefulness to bright and/or closeclusters.The DAML catalogue is a compilation of results fromthe literature. While it is non-homogeneous in nature, it iscurated. The curators choose the best results, when morethan one is available, and keep public logs of what haschanged, of the references to the catalogued parameters anda list of objects that studies have shown not to be real clus-ters as well as the references to those studies. As a compi-lation, it also includes results from the MWSC. Thus, thecomparison with DAML is also a comparison with individualstudies from the literature.The cross-identification of our sample with DAML andwith the MWSC results in 45 and 40 objects in common,respectively. Since DAML also contains values from theMWSC, the later were not included in the DAML plotsto avoid comparing the same points twice. This leaves theDAML comparison sample with 28 clusters. The compar-isons of distance, age and A V are shown in Fig. 6.The A V and distances from both catalogs follow thesame trend as those obtained with our isochrone fits, al-though with some considerable scatter (clearly higher in thecase of the distances from MWSC) and with a tendencyfor smaller catalogued distances for the closer ( < ∼
10 Myr (in the MWSC age scale).In the DAML age comparison we find 4 especially discrepantobjects. They are ESO 332-08, ESO 429-02, NGC 133 andNGC 6885.The cluster sequences for ESO 332-08 and NGC 6885presented in Fig. 7 are well defined and the isochrone fits areclearly adequate. The parameters in DAML for NGC 6885are from Lyng˚a (1988). For ESO 332-08 the parameterswere taken from Kharchenko et al. (2005). We note thatthe same authors later published the MWSC with revisedparameters for ESO 332-08, although the ages coincide inboth catalogues. DAML kept the previous version, whichlisted a larger distance presenting a better fit to CMDs.The isochrone fit in Fig. 7 confirms that the distance inKharchenko et al. (2005) is more accurate than the one inthe MWSC.NGC 133 is the most discrepant cluster in the sample.It is visually identified in DSS images from a small groupof bright stars. The CMD in Fig. 7 displays a bifurcationaround G ∼
16 mag, leading to a redder evolved branch that
MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 Table 2.
Results of mean astrometric parameters obtained using the
Gaia
DR2 stellar proper motion and parallaxes. The meaningof the symbols are as follows: RA IC RS and DE IC RS are the central coordinates of the clusters; r is the radius in which half of theidentified members are located; N is the number of cluster stars; µ α cos δ and µ δ are the proper motion components in mas yr − ; σ isthe dispersion of cluster stars’ proper motions; (cid:36) is the mean parallax of the cluster and σ(cid:36) is the dispersion of the mean parallax. RVand σ RV are the mean and 1 σ dispersion radial velocity obtained for the cluster using Gaia
DR2 data and
N RV is the number of starsused in the determination of RV after outlier rejection.Name RA IC RS DE IC RS r N µ α cos δ σ µ α cos δ µ δ σ µ δ (cid:36) σ (cid:36) RV σ RV N RV ( deg ) ( deg ) ( deg ) ( mas ) ( mas ) ( mas ) ( mas ) ( mas ) ( mas ) ( kms − ) ( kms − )BH 88 151.6211 -51.5557 0.056 89 -6.086 0.377 3.602 0.345 0.386 0.165 22.751 1.100 2Berkeley 64 35.3246 65.8934 0.041 138 -0.551 0.323 0.814 0.408 0.201 0.170Bica 2 308.3153 41.3068 0.060 140 -2.660 0.244 -4.378 0.230 0.555 0.102Bochum 10 160.5040 -59.1324 0.148 264 -7.291 0.317 2.992 0.223 0.378 0.092 0.943 0.290 2Collinder 104 99.1571 4.8155 0.143 179 -1.230 0.411 0.507 0.418 0.514 0.210Czernik 43 351.4483 61.3294 0.057 173 -3.862 0.317 -2.078 0.244 0.350 0.134DC 3 111.7507 -37.5195 0.025 105 -1.214 0.331 2.645 0.472 0.081 0.189Dolidze 1 302.4057 36.5052 0.075 226 -2.721 0.312 -4.961 0.326 0.288 0.118Dolidze 35 291.3465 11.6414 0.064 91 -1.967 0.245 -4.322 0.392 0.288 0.181 22.577 0.174 2ESO 123 26 118.1254 -60.3348 0.105 22 -3.572 0.318 10.909 0.180 1.026 0.044ESO 332 08 253.6906 -40.7299 0.080 201 -0.272 0.316 -1.348 0.313 0.529 0.200ESO 332 13 254.1701 -40.5887 0.058 52 -0.080 0.225 -1.117 0.237 0.558 0.127ESO 392 13 261.7178 -34.7020 0.092 21 1.690 0.262 -2.882 0.176 0.900 0.139ESO 429 02 113.3481 -28.1816 0.050 54 -2.806 0.199 3.673 0.292 0.281 0.155FSR 0224 306.3509 40.2243 0.021 19 -3.242 0.343 -4.373 0.268 0.566 0.068FSR 0236 308.1682 41.4418 0.048 89 -2.491 0.394 -4.076 0.243 0.522 0.163FSR 0377 338.7186 58.3041 0.044 116 -3.219 0.366 -2.155 0.301 0.207 0.159FSR 0441 355.5402 58.5480 0.040 95 -2.049 0.313 -1.160 0.220 0.239 0.169FSR 0591 36.9315 58.7637 0.057 222 -0.231 0.571 -0.566 0.484 0.293 0.214 -72.562 0.864 2FSR 0674 63.0983 48.7296 0.033 49 -0.894 0.419 -0.871 0.243 0.275 0.285FSR 0761 83.3381 39.8388 0.034 85 0.323 0.350 -1.361 0.334 0.253 0.145 -27.299 1.923 2FSR 1443 129.8570 -47.3566 0.054 154 -3.563 0.382 4.147 0.439 0.220 0.171 37.746 1.384 2FSR 1698 230.2346 -59.6270 0.044 161 -4.033 0.341 -3.524 0.307 0.247 0.224Hogg 16 202.2997 -61.2087 0.047 46 -3.479 0.095 -1.645 0.146 0.431 0.065Hogg 22 251.6599 -47.0782 0.044 117 -0.750 0.254 -2.013 0.339 0.343 0.138IC 1442 334.0070 53.9900 0.058 333 -3.083 0.484 -2.884 0.476 0.240 0.198Majaess 65 87.4284 27.0746 0.120 51 -0.258 0.229 -1.063 0.320 0.974 0.160NGC 133 7.8324 63.3583 0.068 284 -2.324 0.431 -0.410 0.250 0.223 0.158 -86.916 0.405 5NGC 1977 83.7945 -4.8018 0.145 93 1.260 0.453 -0.569 0.520 2.590 0.185 27.392 2.361 6NGC 1980 83.8212 -5.9207 0.125 120 1.192 0.388 0.511 0.385 2.583 0.128 25.264 7.055 8NGC 2384 111.2913 -21.0211 0.063 80 -2.303 0.185 3.118 0.220 0.330 0.132NGC 6200 251.0322 -47.4582 0.109 433 -0.950 0.333 -2.244 0.351 0.307 0.265NGC 6444 267.3950 -34.8221 0.059 47 -0.934 0.114 -0.929 0.096 0.521 0.073NGC 6530 271.1088 -24.3572 0.087 80 1.375 0.352 -1.992 0.307 0.762 0.111NGC 6604 274.5127 -12.2449 0.049 88 -0.453 0.208 -2.294 0.314 0.454 0.134NGC 6885 302.9831 26.4935 0.137 726 -3.127 0.356 -5.471 0.413 0.439 0.245 2.378 0.333 4Ruprecht 118 246.1454 -51.9544 0.051 79 -3.152 0.188 -4.345 0.174 0.285 0.107Ruprecht 123 260.7813 -37.8977 0.055 20 1.044 0.172 0.922 0.108 0.604 0.084Ruprecht 55 123.1133 -32.5815 0.064 414 -2.316 0.394 2.921 0.436 0.187 0.174 64.253 1.769 2SAI 43 77.0723 49.8645 0.035 135 0.611 0.390 -0.555 0.389 0.109 0.280Sigma Orionis 84.6860 -2.5959 0.054 45 1.336 0.388 -0.633 0.372 2.479 0.157Stock 3 18.0592 62.3190 0.060 114 -1.895 0.326 -0.357 0.296 0.265 0.132Teutsch 132 77.5140 38.8163 0.057 112 0.326 0.506 -1.536 0.329 0.223 0.234Trapezium-FG 83.8350 -5.4095 0.352 269 1.262 0.449 0.274 0.498 2.557 0.149 23.841 5.161 15vdBergh 130 304.4624 39.3404 0.049 62 -3.609 0.308 -5.075 0.292 0.521 0.154 our isochrone fit follows, but does not include the brightstars. The blue branch does include the brighter stars, whichis what Carraro (2002) identifies as NGC 133 and results inthe parameters listed in DAML. A possibility would be thatwe are looking at different objects along the same line ofsight. However, both branches display the same proper mo-tions and have probable members, which together with thesparseness of the blue branch indicates that the apparentlyyounger sequence is composed of blue stragglers in NGC 133. We conclude that the cluster now revealed by Gaia
DR2 isin fact older than previously estimated.ESO 429-02 is an interesting case. The cluster sequencerevealed by its members is sparse, but still clearly youngwith a pronounced pre-main-sequence (PMS) well fitted bythe lo g t = MNRAS000
DR2 isin fact older than previously estimated.ESO 429-02 is an interesting case. The cluster sequencerevealed by its members is sparse, but still clearly youngwith a pronounced pre-main-sequence (PMS) well fitted bythe lo g t = MNRAS000 , 1–19 (2020) H. Monteiro et al.
Table 3.
Fundamental parameters obtained from the isochrone fits. The last two columns give the distances estimated from parallaxeswith a maximum likelihood estimation assuming a normal distribution and taking into account individual parallax uncertainties. Thestandard errors provided in the distance from parallaxes were estimated considering the calculated 5th and 95th percentile confidenceintervals assuming a symmetric distribution so that σ = r − r /( × . ) , which is equivalent to 1 σ Gaussian uncertainty. The 0.029mas correction (Lindegren et al. 2018) to the mean parallaxes was previously added.Name dist σ dist age σ age [ Fe / H ] σ [ Fe / H ] A V σ A V dist π σ dist π ( pc ) ( pc ) ( dex ) ( dex ) ( dex ) ( dex ) ( mag ) ( mag ) ( pc ) ( pc ) BH 88 2011 321 8.766 0.435 -0.141 0.228 1.612 0.262 1936 1115Berkeley 64 4547 378 8.926 0.046 -0.203 0.171 2.951 0.043 4889 812Bica 2 1550 85 6.746 0.041 0.100 0.214 4.126 0.186 1665 40Bochum 10 2365 19 7.167 0.061 0.179 0.172 1.175 0.066 2390 18Collinder 104 1599 35 7.197 0.081 -0.104 0.171 2.104 0.195 1609 25Czernik 43 2350 113 8.088 0.245 -0.023 0.160 1.931 0.103 2616 74DC 3 7934 607 9.474 0.076 -0.146 0.158 1.042 0.082 8744 3184Dolidze 1 2860 63 7.090 0.032 0.054 0.177 2.049 0.056 2949 95Dolidze 35 2334 98 7.952 0.466 0.204 0.180 3.987 0.050 2603 122ESO 123 26 914 55 8.616 0.089 0.065 0.191 0.451 0.084 948 9ESO 332 08 1693 22 6.911 0.085 0.230 0.187 1.234 0.053 1723 15ESO 332 13 1487 84 6.840 0.139 0.168 0.183 1.380 0.027 1673 51ESO 392 13 1032 85 8.656 0.428 0.074 0.185 1.906 0.201 1057 22ESO 429 02 2875 322 7.113 0.291 -0.120 0.195 1.233 0.108 3141 170FSR 0224 1706 127 6.739 0.106 0.242 0.266 3.061 0.091 1659 29FSR 0236 1610 91 6.877 0.076 0.091 0.198 3.564 0.062 1678 28FSR 0377 3563 297 7.085 0.317 -0.065 0.175 2.159 0.098 4124 162FSR 0441 3473 196 7.079 0.285 -0.153 0.177 2.419 0.039 3579 183FSR 0591 2930 52 7.014 0.182 -0.187 0.205 2.270 0.052 3014 39FSR 0674 2944 656 8.782 0.134 -0.140 0.156 3.107 0.150 3106 193FSR 0761 2485 312 8.770 0.258 -0.112 0.205 1.568 0.368 3107 193FSR 1443 3303 157 8.703 0.416 -0.019 0.159 1.819 0.159 3444 60FSR 1698 3122 118 7.136 0.068 0.228 0.163 2.907 0.035 3341 95Hogg 16 1943 131 7.494 0.262 0.110 0.206 1.422 0.107 2190 68Hogg 22 2354 171 7.076 0.060 0.120 0.170 2.097 0.040 2749 123IC 1442 2710 112 7.665 0.151 -0.100 0.160 1.271 0.277 3378 9075Majaess 65 944 10 8.207 0.167 0.006 0.160 0.768 0.105 945 4NGC 133 3308 311 8.201 0.427 -0.133 0.163 2.310 0.222 3615 142NGC 1977 381 9 6.721 0.064 -0.184 0.170 0.344 0.148 388 27NGC 1980 316 19 6.970 0.049 -0.242 0.175 0.129 0.060 384 18NGC 2384 2494 179 7.318 0.228 -0.147 0.257 0.976 0.098 2775 99NGC 6200 2352 205 7.138 0.060 0.166 0.193 1.858 0.038 2821 152NGC 6444 1492 88 8.632 0.262 0.177 0.191 1.298 0.130 1823 54NGC 6530 1206 39 6.728 0.045 0.373 0.203 1.163 0.037 1265 18NGC 6604 1885 75 6.807 0.118 0.104 0.222 2.804 0.057 2007 59NGC 6885 1453 95 8.092 0.124 0.055 0.192 1.927 0.123 1671 1466Ruprecht 118 2224 125 8.425 0.503 0.386 0.196 1.144 0.041 3004 88Ruprecht 123 1511 74 8.682 0.147 0.188 0.224 1.909 0.169 1622 48Ruprecht 55 4238 286 7.328 0.148 -0.226 0.154 1.639 0.056 4430 43070SAI 43 4451 131 8.410 0.124 -0.198 0.172 1.538 0.075 5009 480Sigma Orionis 303 26 6.997 0.114 -0.092 0.158 0.166 0.040 402 25Stock 3 2747 281 7.226 0.531 -0.100 0.168 2.355 0.073 3051 84Teutsch 132 3474 81 6.992 0.266 -0.160 0.206 2.217 0.069 3567 267Trapezium-FG 381 12 6.778 0.069 -0.146 0.160 0.246 0.089 386 1vdBergh 130 1456 240 6.974 0.091 -0.029 0.222 2.356 0.042 1714 563
Pavani & Bica (2007). Despite the above mentioned limita-tions of these data-sets, their work reveals a CMD that al-though poor, can be plausibly reproduced by an older lo g t = Gaia
DR2 proper motionvector point diagram reveals two over-densities, in which thestronger peak corresponds to the sequence identified in ourwork. It is a possible case of two different objects along thesame line of sight.Of the 11 mean radial velocities of open clusters deter-mined here, 6 are in common with DAML (Bochum 10, NGC6885, Trapezium-FG, NGC 1980, NGC 1977 and Ruprecht 55) published by Dias et al. (2014). The comparison of thissmall sample shows discrepancies ranging from -29 kms − to4 kms − . Considering that the memberships presented in thiswork are superior to those published in Dias et al. (2014), webelieve the radial velocity estimates in this work are morereliable.In the previous sections we validated our cluster pa-rameter determination procedure by comparing distanceswith those from from Gaia
DR2 parallaxes and metallici-ties with those from high resolution spectroscopy. In thissection we confirm that in general, while following the same
MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 trend as those from pre- Gaia studies, our determinations,represent a substantial improvement over the previous val-ues. It is also interesting how the comparisons clearly showthat in this case, a non-homogeneous compilation of param-eters (DAML) can provide a more accurate data-set than anhomogeneously derived catalogue (MWSC). We note, how-ever, that this is seen because we removed the the MWSCvalues from the DAML sample.
We have investigated 45 open clusters with
Gaia
DR2. Fromthe astrometric data (proper motions and parallaxes) we de-termined their stellar membership probabilities, taking intoaccount the full co-variance matrix of the data.For all clusters we estimated mean proper motion andmean parallax considering the member stars (membership ≥ A V were estimated with a new version of the global opti-mization code presented in Monteiro et al. (2017) applied to G BP and G RP photometry using a revised extinction polyno-mial law for Gaia
DR2 and the Galactic abundance gradientas a prior for metallicity. The new procedure was validatedusing a sample of clusters in the literature for which highresolution spectroscopy was available. Our isochrone fittingresults for a high resolution spectroscopy sample are alsopresented. We verify that the PMS portions of the PARSECisochrones fit well the cluster sequences, consistently withthe main sequence fit, indicating that they are suitable foranalyses of young clusters (down to 4 Myr) at least in theGaia photometric bands.This study provides the first determination of dis-tance and age for the cluster Majaess 65 and of age forRuprecht 123. The cluster DC 3 is found to be one of theoldest (5.6 Gyr) and most distant ( ∼ Gaia
DR2 that were alreadyknown clusters listed in DAML. We find that our clusterparameter determinations, represent a substantial improve-ment over the previous values.This work is part of an ongoing project that will bringDAML to the
Gaia era.
DATA AVAILABILITY STATEMENT
ACKNOWLEDGEMENTS
W. S. Dias acknowledges the S˜ao Paulo State AgencyFAPESP (fellowship 2013/01115-6). H. Monteiro would liketo thank FAPEMIG grants APQ-02030-10 and CEX-PPM-00235-12. AM acknowledges the support from the Por-tuguese FCT Strategic Programme UID/FIS/00099/2019for CENTRA. This research was performed using the fa-cilities of the Laborat´orio de Astrof´ısica Computacional daUniversidade Federal de Itajub´a (LAC-UNIFEI).
REFERENCES
Alves J., Bouy H., 2012, A&A, 547, A97Becker W., Fenkart R. B., 1970, in Becker W., Kontopoulos G. I.,eds, IAU Symposium Vol. 38, The Spiral Structure of ourGalaxy. p. 205Bossini D., et al., 2019, A&A, 623, A108Bressan A., Marigo P., Girardi L., Salasnich B., Dal Cero C.,Rubele S., Nanni A., 2012, MNRAS, 427, 127Cantat-Gaudin T., Anders F., 2020, A&A, 633, A99Cantat-Gaudin T., et al., 2018, A&A, 618, A93Cantat-Gaudin T., et al., 2020, A&A, 640, A1Capitanio L., Lallement R., Vergely J. L., Elyajouri M., Monreal-Ibero A., 2017, A&A, 606, A65Carraro G., 2002, A&A, 387, 479Carrera R., et al., 2019, A&A, 623, A80Castelli F., Kurucz R. L., 2003, in Piskunov N., Weiss W. W.,Gray D. F., eds, IAU Symposium Vol. 210, Modelling of Stel-lar Atmospheres. p. A20 ( arXiv:astro-ph/0405087 )Castro-Ginard A., et al., 2020, A&A, 635, A45Chen B., D’Onghia E., Alves J., Adamo A., 2019, arXiv e-prints,p. arXiv:1905.11429Danielski C., Babusiaux C., Ruiz-Dern L., Sartoretti P., ArenouF., 2018, A&A, 614, A19Dias W. S., Alessi B. S., Moitinho A., L´epine J. R. D., 2002, A&A,389, 871Dias W. S., Monteiro H., Caetano T. C., L´epine J. R. D., AssafinM., Oliveira A. F., 2014, A&A, 564, A79Dias W. S., Monteiro H., L´epine J. R. D., Prates R., GneidingC. D., Sacchi M., 2018, MNRAS, 481, 3887Dias W. S., Monteiro H., L´epine J. R. D., Barros D. A., 2019,MNRAS, 486, 5726Donor J., et al., 2020, AJ, 159, 199Fitzpatrick E. L., Massa D., Gordon K. D., Bohlin R., ClaytonG. C., 2019, ApJ, 886, 108Gaia Collaboration et al., 2016, A&A, 595, A1Gaia Collaboration et al., 2018a, A&A, 616, A1Gaia Collaboration et al., 2018b, A&A, 616, A10Janes K., Adler D., 1982, ApJS, 49, 425Kharchenko N. V., Piskunov A. E., R¨oser S., Schilbach E., ScholzR. D., 2005, A&A, 438, 1163Kharchenko N. V., Piskunov A. E., Schilbach E., R¨oser S., ScholzR.-D., 2013, A&A, 558, A53Kounkel M., Covey K., 2019, AJ, 158, 122Krone-Martins A., Moitinho A., 2014, A&A, 561, A57Lindegren L., et al., 2018, A&A, 616, A2Liu L., Pang X., 2019, ApJS, 245, 32Luri X., et al., 2018, A&A, 616, A9MNRAS000
Alves J., Bouy H., 2012, A&A, 547, A97Becker W., Fenkart R. B., 1970, in Becker W., Kontopoulos G. I.,eds, IAU Symposium Vol. 38, The Spiral Structure of ourGalaxy. p. 205Bossini D., et al., 2019, A&A, 623, A108Bressan A., Marigo P., Girardi L., Salasnich B., Dal Cero C.,Rubele S., Nanni A., 2012, MNRAS, 427, 127Cantat-Gaudin T., Anders F., 2020, A&A, 633, A99Cantat-Gaudin T., et al., 2018, A&A, 618, A93Cantat-Gaudin T., et al., 2020, A&A, 640, A1Capitanio L., Lallement R., Vergely J. L., Elyajouri M., Monreal-Ibero A., 2017, A&A, 606, A65Carraro G., 2002, A&A, 387, 479Carrera R., et al., 2019, A&A, 623, A80Castelli F., Kurucz R. L., 2003, in Piskunov N., Weiss W. W.,Gray D. F., eds, IAU Symposium Vol. 210, Modelling of Stel-lar Atmospheres. p. A20 ( arXiv:astro-ph/0405087 )Castro-Ginard A., et al., 2020, A&A, 635, A45Chen B., D’Onghia E., Alves J., Adamo A., 2019, arXiv e-prints,p. arXiv:1905.11429Danielski C., Babusiaux C., Ruiz-Dern L., Sartoretti P., ArenouF., 2018, A&A, 614, A19Dias W. S., Alessi B. S., Moitinho A., L´epine J. R. D., 2002, A&A,389, 871Dias W. S., Monteiro H., Caetano T. C., L´epine J. R. D., AssafinM., Oliveira A. F., 2014, A&A, 564, A79Dias W. S., Monteiro H., L´epine J. R. D., Prates R., GneidingC. D., Sacchi M., 2018, MNRAS, 481, 3887Dias W. S., Monteiro H., L´epine J. R. D., Barros D. A., 2019,MNRAS, 486, 5726Donor J., et al., 2020, AJ, 159, 199Fitzpatrick E. L., Massa D., Gordon K. D., Bohlin R., ClaytonG. C., 2019, ApJ, 886, 108Gaia Collaboration et al., 2016, A&A, 595, A1Gaia Collaboration et al., 2018a, A&A, 616, A1Gaia Collaboration et al., 2018b, A&A, 616, A10Janes K., Adler D., 1982, ApJS, 49, 425Kharchenko N. V., Piskunov A. E., R¨oser S., Schilbach E., ScholzR. D., 2005, A&A, 438, 1163Kharchenko N. V., Piskunov A. E., Schilbach E., R¨oser S., ScholzR.-D., 2013, A&A, 558, A53Kounkel M., Covey K., 2019, AJ, 158, 122Krone-Martins A., Moitinho A., 2014, A&A, 561, A57Lindegren L., et al., 2018, A&A, 616, A2Liu L., Pang X., 2019, ApJS, 245, 32Luri X., et al., 2018, A&A, 616, A9MNRAS000 , 1–19 (2020) H. Monteiro et al.
Figure 7.
CMDs and isochrone fits to the Gaia DR2 data for the clusters investigated in this study. Probable member stars are shownin blue dots, with more intense tones indicating higher membership probability. The light-gray dots mark non-member stars in the field.MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 Figure 7.
CMDs and isochrone fits (continued)MNRAS000
CMDs and isochrone fits (continued)MNRAS000 , 1–19 (2020) H. Monteiro et al.
Figure 7.
CMDs and isochrone fits (continued) MNRAS , 1–19 (2020) arameters for 45 open clusters with Gaia DR2 Figure 7.
CMDs and isochrone fits (continued)Lyng˚a G., 1988, in European Southern Observatory Conferenceand Workshop Proceedings. pp 379–382Lyra W., Moitinho A., van der Bliek N. S., Alves J., 2006, A&A,453, 101Ma´ız Apell´aniz J., Weiler M., 2018, A&A, 619, A180Maurya J., Joshi Y. C., Gour A. S., 2020, arXiv e-prints, p.arXiv:2005.07375Moitinho A., 2010, in de Grijs R., L´epine J. R. D., eds, IAU Sym-posium Vol. 266, Star Clusters: Basic Galactic Building BlocksThroughout Time and Space. pp 106–116 ( arXiv:0911.1459 ),doi:10.1017/S1743921309990949Monteiro H., Dias W. S., 2019, MNRAS, 487, 2385Monteiro H., Dias W. S., Hickel G. R., Caetano T. C., 2017,New Astron., 51, 15Netopil M., Paunzen E., Carraro G., 2015, A&A, 582, A19 Netopil M., Paunzen E., Heiter U., Soubiran C., 2016, A&A, 585,A150Pavani D. B., Bica E., 2007, A&A, 468, 139Perren G. I., Giorgi E. E., Moitinho A., Carraro G., Pera M. S.,V´azquez R. A., 2020, A&A, 637, A95Roeser S., Demleitner M., Schilbach E., 2010, AJ, 139, 2440Sim G., Lee S. H., Ann H. B., Kim S., 2019, Journal of KoreanAstronomical Society, 52, 145Skrutskie M. F., et al., 2006, AJ, 131, 1163Soubiran C., et al., 2018, A&A, 619, A155The Astropy Collaboration et al., 2018, AJ, 156, 123Wang S., Chen X., 2019, ApJ, 877, 116MNRAS000
CMDs and isochrone fits (continued)Lyng˚a G., 1988, in European Southern Observatory Conferenceand Workshop Proceedings. pp 379–382Lyra W., Moitinho A., van der Bliek N. S., Alves J., 2006, A&A,453, 101Ma´ız Apell´aniz J., Weiler M., 2018, A&A, 619, A180Maurya J., Joshi Y. C., Gour A. S., 2020, arXiv e-prints, p.arXiv:2005.07375Moitinho A., 2010, in de Grijs R., L´epine J. R. D., eds, IAU Sym-posium Vol. 266, Star Clusters: Basic Galactic Building BlocksThroughout Time and Space. pp 106–116 ( arXiv:0911.1459 ),doi:10.1017/S1743921309990949Monteiro H., Dias W. S., 2019, MNRAS, 487, 2385Monteiro H., Dias W. S., Hickel G. R., Caetano T. C., 2017,New Astron., 51, 15Netopil M., Paunzen E., Carraro G., 2015, A&A, 582, A19 Netopil M., Paunzen E., Heiter U., Soubiran C., 2016, A&A, 585,A150Pavani D. B., Bica E., 2007, A&A, 468, 139Perren G. I., Giorgi E. E., Moitinho A., Carraro G., Pera M. S.,V´azquez R. A., 2020, A&A, 637, A95Roeser S., Demleitner M., Schilbach E., 2010, AJ, 139, 2440Sim G., Lee S. H., Ann H. B., Kim S., 2019, Journal of KoreanAstronomical Society, 52, 145Skrutskie M. F., et al., 2006, AJ, 131, 1163Soubiran C., et al., 2018, A&A, 619, A155The Astropy Collaboration et al., 2018, AJ, 156, 123Wang S., Chen X., 2019, ApJ, 877, 116MNRAS000 , 1–19 (2020) H. Monteiro et al.
APPENDIX A: RESULTS FOR THE HIGH RESOLUTION SPECTROSCOPY VALIDATION SAMPLE
MNRAS , 1–19 (2020) a r a m e t e r s f o r e n c l u s t e r s w i t h G a i a D R Table A1: Results of isochrone fits done using the [ Fe / H ] prior based on the metallicity gradient from Donor et al. (2020) for the control sample described in section 4.2.The available [ Fe / H ] values and uncertainties from the literature used in the comparisons are also presented. The h and l suffixes in values from Netopil et al. (2016)denote their high and low quality samples, respectively. Carrera et al. (2019) Netopil et al. (2016)Name dist σ dist a g e σ age A V σ A V [ Fe / H ] σ [ Fe / H ] [ Fe / H ] σ [ Fe / H ] [ Fe / H ] h σ [ Fe / H ] h [ Fe / H ] l σ [ Fe / H ] l ASCC 21 344 2 7.062 0.039 0.201 0.029 -0.026 0.165 0.01 0.09 - - - -Basel 11b 1663 71 8.451 0.068 1.856 0.094 -0.090 0.184 0.01 0.05 - - - -Berkeley 17 3278 105 9.791 0.110 1.923 0.090 -0.173 0.157 -0.10 0.04 -0.06 - - -Berkeley 19 6393 790 9.271 0.042 1.493 0.134 -0.431 0.187 -0.22 - - - - -Berkeley 31 7019 329 9.502 0.027 0.540 0.032 -0.302 0.157 -0.31 0.04 - - - -Berkeley 33 4467 277 8.520 0.051 2.085 0.043 -0.268 0.181 -0.23 0.11 - - -0.26 0.05Berkeley 43 1994 93 7.660 0.338 4.774 0.047 0.171 0.156 0.00 - - - - -Berkeley 53 4525 263 8.885 0.021 4.391 0.044 -0.090 0.171 -0.02 0.03 - - - -Berkeley 66 8738 1318 8.637 0.088 3.886 0.072 -0.110 0.180 -0.12 0.01 - - - -Berkeley 71 3203 138 8.827 0.030 3.045 0.052 -0.100 0.157 -0.20 0.03 - - - -Berkeley 9 1720 135 9.187 0.060 2.976 0.037 -0.100 0.198 -0.17 0.18 - - - -Berkeley 98 3391 78 9.504 0.032 0.753 0.022 -0.090 0.154 0.03 0.02 - - - -Collinder 69 398 1 6.880 0.043 0.405 0.045 -0.100 0.160 -0.01 0.06 - - - -Collinder 95 661 8 6.791 0.152 0.886 0.643 -0.065 0.176 -0.03 0.02 - - - -Czernik 21 3900 510 8.915 0.123 3.211 0.160 -0.268 0.180 -0.24 0.01 - - - -Czernik 23 3070 172 8.474 0.544 1.761 0.104 -0.100 0.177 -0.25 - - - - -Czernik 30 5729 365 9.466 0.023 0.976 0.098 -0.289 0.213 -0.28 0.02 - - - -FSR 0496 1506 70 8.814 0.023 3.112 0.054 -0.130 0.166 -0.07 - - - - -FSR 0542 5506 709 8.889 0.089 3.514 0.107 -0.177 0.290 -0.19 - - - - -FSR 0667 1100 28 8.655 0.129 1.502 0.167 0.154 0.227 0.03 0.01 - - - -FSR 0716 3388 169 9.043 0.051 1.242 0.122 -0.243 0.183 -0.30 - - - - -FSR 0941 4029 37 8.826 0.085 2.445 0.057 -0.100 0.225 -0.23 - - - - -FSR 0942 3151 510 8.840 0.139 2.414 0.178 -0.122 0.172 -0.28 - - - - -Gulliver 6 415 2 7.137 0.079 0.304 0.077 0.031 0.183 -0.10 - - - - -Haffner 4 3758 260 8.950 0.118 1.430 0.107 -0.326 0.191 -0.13 - - - - -IC 1369 2683 284 7.773 0.695 2.594 0.122 -0.018 0.159 -0.02 0.01 - - - -IC 1805 1849 113 6.805 0.081 2.296 0.022 -0.056 0.189 0.32 - - - - -King 15 2727 140 8.493 0.452 1.926 0.083 -0.100 0.243 -0.05 - - - - -Kronberger 57 2211 663 6.738 0.799 4.336 0.148 0.383 0.308 0.02 - - - - -Melotte 20 174 3 7.858 0.025 0.386 0.040 0.036 0.162 0.08 0.07 0.14 0.11 - -Melotte 22 136 1 8.090 0.097 0.154 0.051 0.127 0.167 0.06 0.08 -0.01 0.05 - -Melotte 71 1966 47 9.097 0.031 0.512 0.060 -0.100 0.176 -0.09 0.02 -0.27 - - -NGC 1193 5166 206 9.713 0.057 0.674 0.028 -0.221 0.159 -0.25 0.01 -0.22 - - -NGC 1245 2636 42 9.096 0.016 0.871 0.026 -0.100 0.153 -0.06 0.03 0.02 0.03 -0.05 0.06NGC 136 4648 282 8.376 0.681 2.124 0.133 -0.124 0.172 -0.22 - - - - -NGC 1664 1197 23 8.790 0.022 0.918 0.068 -0.127 0.156 -0.01 - - - - -NGC 1798 4741 243 9.139 0.014 1.725 0.086 -0.294 0.191 -0.18 0.02 - - - -NGC 1817 1544 39 9.078 0.017 0.785 0.067 -0.119 0.166 -0.09 - -0.11 0.03 -0.16 0.03NGC 1857 2506 114 8.377 0.398 1.679 0.087 -0.192 0.176 -0.12 - - - - - M N R A S , ( ) H . M o n t e i r o e t a l . Table A1: continued from previous pageCarrera et al. (2019) Netopil et al. (2016)Name dist σ dist a g e σ age A V σ A V [ Fe / H ] σ [ Fe / H ] [ Fe / H ] σ [ Fe / H ] [ Fe / H ] h σ [ Fe / H ] h [ Fe / H ] l σ [ Fe / H ] l NGC 188 1836 5 9.789 0.018 0.353 0.072 -0.062 0.161 0.13 0.05 0.11 0.04 -0.02 0.09NGC 1907 1539 54 8.681 0.142 1.672 0.147 -0.268 0.174 -0.05 0.01 - - - -NGC 1912 1058 22 8.479 0.123 0.937 0.068 0.048 0.164 -0.07 0.02 -0.10 0.14 - -NGC 2158 4030 306 9.381 0.065 1.495 0.067 -0.268 0.156 -0.15 0.03 - - -0.32 0.08NGC 2168 845 16 8.145 0.168 0.903 0.077 -0.110 0.173 -0.13 0.07 - - -0.21 0.10NGC 2183 786 34 7.006 0.202 1.670 0.473 -0.100 0.199 -0.08 0.08 - - - -NGC 2243 4005 106 9.542 0.044 0.168 0.024 -0.358 0.154 -0.42 - - - -0.50 0.08NGC 2244 1287 107 7.093 0.143 1.586 0.091 -0.121 0.165 -0.23 0.09 - - - -NGC 2304 3814 143 8.977 0.034 0.308 0.103 -0.275 0.171 -0.09 0.09 - - - -NGC 2318 1271 41 8.878 0.076 0.839 0.128 0.078 0.158 0.01 - - - - -NGC 2324 3732 70 8.749 0.036 0.814 0.073 -0.215 0.156 -0.15 0.05 -0.22 0.07 - -NGC 2355 1837 20 9.086 0.034 0.329 0.015 0.042 0.153 -0.11 - -0.05 0.08 -0.08 0.08NGC 2420 2471 101 9.407 0.045 0.123 0.009 -0.218 0.158 -0.12 0.03 -0.05 0.02 -0.21 0.09NGC 2682 855 4 9.561 0.004 0.185 0.030 -0.031 0.154 0.02 0.07 0.03 0.05 0.00 0.06NGC 6705 1922 50 8.440 0.163 1.502 0.071 0.046 0.16 0.16 0.03 0.12 0.09 0.25 0.05NGC 6791 4422 74 9.946 0.053 0.391 0.052 0.221 0.165 0.40 0.07 0.42 0.05 0.35 0.07NGC 6811 1097 17 9.021 0.022 0.232 0.048 0.015 0.161 -0.01 0.04 0.03 0.01 - -NGC 6819 2310 141 9.459 0.037 0.507 0.026 0.011 0.166 0.10 0.04 0.09 0.01 -0.04 0.08NGC 6866 1392 44 8.844 0.034 0.477 0.076 0.183 0.156 0.04 0.02 - - - -NGC 7058 362 2 7.860 0.584 0.291 0.183 -0.100 0.175 0.12 0.04 - - - -NGC 7062 2109 168 8.643 0.397 1.730 0.154 -0.081 0.173 0.04 - - - - -NGC 752 444 4 9.179 0.019 0.166 0.061 -0.037 0.159 0.01 - -0.03 0.06 -0.09 0.13Teutsch 12 3939 331 8.948 0.036 1.969 0.118 -0.118 0.185 -0.14 0.02 - - - -Teutsch 51 5387 443 8.817 0.068 3.311 0.083 -0.285 0.181 -0.28 0.04 - - - -Tombaugh 4 3127 215 8.918 0.062 3.218 0.065 -0.103 0.154 -0.47 - - - - -Trumpler 26 1336 76 8.669 0.114 1.703 0.108 0.175 0.172 0.28 0.05 - - - -Trumpler 3 663 8 8.094 0.079 0.931 0.042 0.156 0.150 -0.22 - - - - -Trumpler 5 3275 56 9.536 0.025 1.846 0.066 -0.152 0.164 -0.36 0.02 -0.44 0.07 -0.47 0.05 M N R A S , ( ) arameters for 45 open clusters with Gaia DR2 [!H] Table B1.
Removed clusters. Central coordinates and radii from DAML.Name RA J DE J radius ( deg ) ( deg ) ( deg ) ASCC 94 273.9000 -14.9900 0.250BH 4 114.4333 -36.0667 0.017Bochum 1 96.3542 19.7667 0.217Collinder 347 266.5750 -29.3333 0.083Collinder 92 95.7250 5.1167 0.092Dolidze 13 12.4250 64.1264 0.133Dolidze 24 101.1708 1.6847 0.157Dolidze 35 291.3500 11.6583 0.058Dolidze 41 304.7042 37.7500 0.092Dolidze 49 101.7667 -0.0069 0.018ESO 522 05 273.2208 -24.3639 0.037FSR 0182 297.9417 33.5119 0.010FSR 0258 311.2083 43.9150 0.013FSR 0354 332.8000 57.6994 0.043FSR 0453 356.8542 63.2264 0.037FSR 0522 13.4583 65.7933 0.006FSR 0717 71.5250 42.1342 0.018FSR 0891 94.3708 22.4272 0.012FSR 0929 96.3833 17.7200 0.007FSR 1535 151.9792 -59.1969 0.018Hogg 11 167.9042 -60.4000 0.017Kronberger 39 163.5583 -61.7378 0.007Majaess 50 71.3625 41.9758 0.142Majaess 95 124.4708 -35.8800 0.025NGC 2013 86.0042 55.7933 0.050Patchick 78 8.2917 65.1167 0.013Ruprecht 120 248.7917 -48.2833 0.025Ruprecht 136 269.8250 -24.7000 0.025Ruprecht 59 124.8375 -34.4833 0.025Teutsch 64 128.1292 -41.9881 0.038
APPENDIX B: REMOVED CLUSTERS
Here we identify the clusters that have been removed from our studied sample, as discussed in section 5.
This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS000