Investigating the age and structure of the infrared old open clusters LK1, LK10, FSR1521 and FSR1555
aa r X i v : . [ a s t r o - ph ] O c t Mon. Not. R. Astron. Soc. , 1–15 (0000) Printed 31 October 2018 (MN L A TEX style file v2.2)
Investigating the age and structure of the infrared oldopen clusters LK 1, LK 10, FSR 1521 and FSR 1555
C. Bonatto ⋆ and E. Bica † Departamento de Astronomia, Universidade Federal do Rio Grande do SulAv. Bento Gon¸calves 9500, Porto Alegre 91501-970, RS, Brazil
31 October 2018
ABSTRACT
The combination of several mass-decreasing processes may critically affect the struc-ture of open clusters (OCs), to the point that most dissolve into the field in a time-scaleshorter than ≈ J , H , and K s bands, and the in-trinsic morphologies of the target OCs within these diagrams are revealed by applyinga field-star decontamination algorithm. Fundamental parameters are estimated withPadova isochrones built for the 2MASS filters. We derive extinctions to the objectswithin the range 3 . A V .
9, which makes them suitable for the near-infraredanalysis, ages within 1 . − . . − . . . ≈ χ = 2 . ± .
4) when compared to the Salpeter value ( χ = 1 . m ( M ⊙ ) . M ⊙ , respectively.The mass estimated in the restricted magnitude range for the remaining, more distantOCs is within 260 m ( M ⊙ ) Key words: (Galaxy:) open clusters and associations: general; (Galaxy:) open clus-ters and associations: individual:LK 1, LK 10, FSR 1521, and FSR 1555.
The majority of the Galactic open clusters (OCs), the lessmassive ones in particular, do not survive the ≈ ⋆ E-mail: [email protected] † E-mail: [email protected] mass loss associated with stellar evolution, mass segrega-tion and evaporation, tidal interactions with the Galacticdisc and bulge, and collisions with giant molecular clouds.As clusters age, these mechanisms accelerate the internaldynamical evolution, which leads to important changes inthe structure and flat, eroded mass functions. In mostcases, OCs have their stellar content completely dissolved inthe Galactic field, or leave only poorly-populated remnants(Pavani & Bica 2007 and references therein).Ample evidence gathered on theoretical (e.g. c (cid:13) C. Bonatto and E. Bica
Spitzer 1958; Lamers & Gieles 2006), N-body (e.g.Baumgardt & Makino 2003; Goodwin & Bastian 2006;Khalisi, Amaro-Seoane & Spurzem 2007), and observa-tional (e.g. van den Bergh 1957; Oort 1958; von Hoerner1958; Piskunov et al. 2007) grounds indicate that, nearthe Solar circle, the disruption-time scale ( t dis ) is shorterthan ∼ t dis ∼ M . (Lamers & Gieles 2006). Thus, for clusters with mass within10 − M ⊙ , a disruption time of 75 . t dis ( Myr ) . R GC .
150 pc) Galactic tidal fields can dissolvea massive star cluster in a time-scale as short as ∼
50 Myr(Portegies Zwart et al. 2002).What should be expected from the above scenariois that only a small fraction of the OCs survive theGyr age-barrier, with the successful ones spending mostof their existences preferentially at large Galactocentricdistances. Indeed, present-day statistics show that of the ≈ database, 180 are older than 1 Gyr, and only 18 ( ≈ J , H , and K s photometryobtained from the 2MASS Point Source Catalogue (PSC).The spatial and photometric uniformity of 2MASS, whichallow extraction of large surrounding fields that provide The Two Micron All Sky Survey, All Skydata release (Skrutskie et al. 1997), available at
Table 1.
Previous identifications of old clusters by our groupCluster Age Reference(Gyr)FSR 869 1.5 Bonatto & Bica (2008c)FSR 942 1.0 Bonatto & Bica (2008c)FSR 70 & & & α (2000) =05 h m s and δ (2000) = +39 ◦ ′ ′′ , is probably physi-cally associated with the planetary nebula PK 167-0.1. high star-count statistics, make it an excellent resource togather photometric data on a broad variety of star clus-ters, the wide field ones in particular. For this purpose wehave developed quantitative tools to statistically disentan-gle cluster evolutionary sequences from field stars in colour-magnitude diagrams (CMDs), which are subsequently usedto investigate the nature of star cluster candidates and toderive astrophysical parameters of the confirmed clusters(e.g. Bica, Bonatto & Camargo 2008). Basically, we apply (i) field-star decontamination to quantify the statistical sig-nificance of the CMD morphology, which is fundamental toderiving reddening, age, and distance from the Sun, and (ii) colour-magnitude filters, which are essential for intrin-sic stellar radial density profiles (RDPs), as well as lumi-nosity and mass functions (MFs). In particular, the use offield-star decontamination in the construction of CMDs hasproved to constrain the age and distance more than whenworking with the raw (observed) photometry, especially forlow-latitude OCs (Bonatto et al. 2006a).This paper is organised as follows. In Sect. 2 we recallrecent additions to the known old OCs made by our group.Sect. 3 contains basic properties and reviews literature data(where available) on the present star cluster candidates. InSect. 4 we present the 2MASS photometry, build CMDs, andapply the field-star decontamination algorithm. In Sect. 5we derive cluster fundamental parameters. Sect. 6 describescluster structure by means of stellar RDPs. In Sect. 7 weprovide estimates of cluster mass. In Sect. 8 we compare thestructural parameters and dynamical state of the presentclusters with those of a sample of nearby OCs, we also dis-cuss effects of the location in the Galaxy on their structure.Concluding remarks are given in Sect. 9. c (cid:13) , 1–15 nvestigation of 4 old infrared OCs Figure 1.
Top panels: 5 ′ × ′ K s images of LK 1 (left) and LK 10 (right). Bottom: same for FSR 1521 (left) and FSR 1555(right). Images provided by the 2MASS Image Service. The small circle indicates the central coordinates (cols. 4 and 5 of Table 2).Figure orientation: North to the top and East to the left. In recent years our group has been systematically analysinginfrared clusters or candidates, establishing their nature andderiving cluster fundamental parameters with the 2MASScatalogue (e.g. Bica, Bonatto & Camargo 2008). We make use of a field decontamination algorithm (described inSect. 4.1) to statistically extract estimated cluster sequencesfrom CMDs. Especially for crowded fields, the cluster se-quence isolation requires some form of membership selection(e.g. Bonatto & Bica 2007b). Old OCs are intrinsically rarerthan young clusters (Sect. 1), but the 2MASS catalogue,coupled to the statistical tools that we have developed to c (cid:13) , 1–15 C. Bonatto and E. Bica
Table 2.
Previous data and present results on the clusters
Literature This paperCluster α (2000) δ (2000) α (2000) δ (2000) ℓ b Age AV d ⊙ R GC X GC Y GC Z GC(hms) ( ◦ ′ ′′ ) (hms) ( ◦ ′ ′′ ) ( ◦ ) ( ◦ ) (Gyr) (mag) (kpc) (kpc) (kpc) (kpc) (kpc)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)LK 1 20:24:25 +36:24:18 ( † ) +36:24:30.0 75.24 − .
69 1 . ± . . ± . . ± . . ± . − . ± . . ± . − . ± . † ) +39:58:43.2 79.84 − .
92 1 . ± . . ± . . ± . . ± . − . ± . . ± . − . ± . − − − .
64 2 . ± . . ± . . ± . . ± . − . ± . − . ± . − . ± . − − .
16 1 . ± . . ± . . ± . . ± . − . ± . − . ± . . ± . † ): same value as in the literature; Col. 9: reddening towards the cluster’s central region (Sect. 5). Col. 10: distance from the Sun.Col. 11: cluster Galactocentric distance for R ⊙ = 7 . deal with clusters and comparison fields allow the detection(and derivation of reliable astrophysical parameters) of olderOCs (e.g. Bonatto & Bica 2007b; Bica & Bonatto 2008).Recently, Froebrich, Scholz & Raftery (2007) provideda catalogue of star cluster candidates corresponding to 1021stellar overdensities detected in the 2MASS database. Thiscatalogue covers | b | < ◦ and all Galactic longitudes,and has become an important source of new star clus-ters. Several followup studies have explored the FSR cat-alogue with different approaches, revealing new globularclusters, such as FSR 1735 (Froebrich, Meusinger & Scholz2007) and FSR 1767 (Bonatto et al. 2007), and theprobable GCs FSR 584 (Bica et al. 2007) and FSR 190(Froebrich, Meusinger & Davis 2007). We show in Table 114 recent identifications of old OCs made by our group.Three of them are optical objects, while the remaining 11are infrared ones. Compared to the DAML02 optical cata-logue of OCs, our results increased the known sample of oldOCs by ≈ Le Duigou & Kn¨odlseder (2002) provided a list of 17 starclusters and candidates in the Cygnus direction. As theypoint out, 12 of these objects (hereafter designated by LK)had already been found by Dutra & Bica (2001). Most ofthe LK objects appear to be embedded clusters, but someof them, like LK 1 and LK 10, have features typical of olderclusters. With 2MASS photometry, they were able to es-timate some cluster parameters. For LK 1 they derived aradius (containing 90% of the stars) R = 3 . ′
0, the dis-tance modulus DM = 11 .
0, the absorption in the K band A K = 1 . − .
0, and a mass within M = 1100 − M ⊙ ;they suggested that LK 1 may be a rather evolved clus-ter. As for LK 10, they found R = 6 . ′ DM = 11 . A K = 0 . − . M = 1010 − M ⊙ , a rather steep massfunction (MF: φ ( m ) ∝ m − (1+ χ ) ) slope χ = 1 . ± .
25, andsuggested that it is quite evolved. Near-infrared 2MASS K s images of LK 1 and LK 10, covering 5 ′ × ′ fields, are shownin Fig. 1 (top panels).FSR 1521 was classified by Froebrich, Scholz & Raftery (2007) as a highly probable star cluster candidate. Theyderived the core and tidal radii (measured in 2MASS H images) R Hc = 1 . ′ R Ht = 25 . ′
3, respectively. FSR 1521can be seen in the 5 ′ × ′ K s image shown in Fig. 1 (bottom-left panel).FSR 1555 was also classified as a highly probable starcluster candidate by Froebrich, Scholz & Raftery (2007),who derived R Hc = 1 . ′ R Ht = 11 . ′ ′ × ′ K s image shown in Fig. 1(bottom-right panel). In general, the clusters in the presentsample are very contaminated by field stars, which requiresspecific tools to analyse them.Table 2 provides fundamental data on the objects,where the literature coordinates are given in cols. 2 and3. However, when we built the RDPs based on these co-ordinates (Sect. 6), we noticed that, in all cases, the coor-dinates where the maximum stellar number-density occursare slightly shifted with respect to the literature positions.Thus, hereafter we will refer as cluster coordinates those thatmaximise the central stellar density (given in cols. 4-7). Theage, central reddening, distance from the Sun, Galactocen-tric distance, and the components projected onto the Galac-tic plane derived in the present study (Sect. 5) are given inCols. 8 to 14. Photometry in the 2MASS J , H , and K s bands was ex-tracted in circular fields of radius R ext centred on the co-ordinates of the objects (Table 2) by means of VizieR . R ext should be large enough to allow the determinationof the background level (Sect. 6). In the present cases, R ext = 30 ′ (FSR 1521 and FSR 1555), R ext = 40 ′ (LK 1),and R ext = 60 ′ (LK 10), which are considerably larger thanthe cluster radius (Sect. 6 and col. 5 of Table 4). Previ-ous works with OCs in different environments (Sect. 1) haveshown that in the absence of a populous neighbouring clus-ter and significant differential absorption (Sect. 4.1.1), wideextraction areas provide the necessary statistics for a con-sistent colour and magnitude characterisation of the fieldstars. For decontamination purposes, comparison fields wereextracted within wide rings located beyond the cluster radii.As photometric quality constraint, the 2MASS extractionswere restricted to stars (i) brighter than those of the 99.9% http://vizier.u-strasbg.fr/viz-bin/VizieR?-source=II/246 c (cid:13) , 1–15 nvestigation of 4 old infrared OCs LK1: Raw-20 -10 0 10 20 ∆α -20 -10 0 10 20 ∆δ σ LK1: Clean-20 -10 0 10 20 ∆α -20 -10 0 10 20 ∆δ σ LK10: Raw-30 -20 -10 0 10 20 30 ∆α -30-20-10 0 10 20 30 ∆δ σ LK10: Clean-30 -20 -10 0 10 20 30 ∆α -30-20-10 0 10 20 30 ∆δ σ Figure 2.
Stellar surface-density σ (stars arcmin − ) of LK 1 (leftpanels) and LK 10 (right). The curves were computed for a meshsize of 3 ′ × ′ , centred on the coordinates in Table 2. The observed(raw) and field-star decontaminated photometry are shown in thetop and bottom panels, respectively. Point Source Catalogue completeness limit in the clusterdirection, and (ii) with errors in J , H , and K s smaller than0.3 mag. The 99.9% completeness limits refer to field stars,and depend on Galactic coordinates. For the present clus-ters, the fraction of stars with J , H , and K s uncertaintiessmaller than 0.06 mag is ≈ A J /A V = 0 . A H /A V = 0 . A K S /A V =0 . A J = 2 . × E ( J − H ) (Dutra, Santiago & Bica2002), for a constant total-to-selective absorption ratio R V = 3 .
1. These ratios were derived from the extinctioncurve of Cardelli, Clayton & Mathis (1989).CMDs displaying the J × ( J − H ) and J × ( J − K s )colours built with the raw photometry of the present clus-ters are shown in Figs. 4 - 7 (top panels). For all clusters, thesampled region is larger than the respective core (Table 4).When qualitatively compared with the CMDs extractedfrom the equal-area comparison fields (middle panels), fea-tures typical of old OCs are apparent. A relatively populousred clump (at 13 . . J . . . . ( J − H ) . .
5) anda main-sequence turn off (MSTO) stand out over the fieldcontamination of LK 1 (Fig. 4). A somewhat less-populousred clump (11 . . J . .
7, 1 . . ( J − H ) . .
5) andabout 3 MS mags ( J & .
4) are seen in LK 10 (Fig. 5). LK 1(Fig. 4), FSR 1521 (Fig. 6), and FSR 1555 (Fig. 7) presentsimilar CMDs, with clear red clumps, the MSTO, and about1 mag of the MS below. The red clump of FSR 1521 occurs According to the 2MASS Level 1 Requirement, at
FSR1555: Raw-20 -10 0 10 20 ∆α -20 -10 0 10 20 ∆δ σ FSR1555: Clean-20 -10 0 10 20 ∆α -20 -10 0 10 20 ∆δ σ FSR1521: Raw-20 -10 0 10 20 ∆α -20-10 0 10 20 ∆δ σ FSR1521: Clean-20 -10 0 10 20 ∆α -20-10 0 10 20 ∆δ σ Figure 3.
Same as Fig. 2 for FSR 1555 (left) and FSR 1521(right). at 12 . . J . . . . ( J − H ) . .
0, while forFSR 1555 it is at 12 . . J . . . . ( J − H ) . . σ (inunits of stars arcmin − ), in a rectangular mesh with cellsof dimensions 3 ′ × ′ . The meshes reach total offsets of | ∆ α | = | ∆ δ | ≈ ′ with respect to the centre (Table 2),in right ascension and declination; for LK 10 we use offsetsof 30 ′ . In all clusters, the core (Table 4) is contained in thecentral cell.With respect to the surface-densities built with the ob-served (raw) photometry (top panels of Figs. 2 and 3), animportant excess appears in the central cell, except for LK 10which, because of the contamination by disc stars, presents arather irregular distribution. FSR 1521 and FSR 1555, on theother hand, clearly detach in the central cell (Fig. 3) againstmore uniform surrounding fields. As shown in the bottompanels, the cluster overdensities are clearly enhanced withrespect to the surroundings in the field-star decontaminatedsurfaces (Sect. 4.1). As expected of low-latitude clusters (Table 2), the stellarsurface-density in the direction of the objects (Figs. 2 and 3)clearly shows that field-star contamination, essentially fromdisc stars, should be taken into account. This fact is con-firmed by the qualitative comparison between the CMDs ex-tracted within the cluster and in the field (Figs. 4-7). Thus,the field-star contribution should be quantified in each caseto better define the intrinsic CMD morphology.Field-star decontamination is a very important, yet dif-ficult, step in the identification and characterisation of starclusters. Several approaches have been used to this purpose(e.g. Mercer et al. 2005), and most of them are based essen- c (cid:13) , 1–15 C. Bonatto and E. Bica S )LK1 Raw (R=2‘)Field stars Same areaClean Figure 4.
R < ′ regionof LK 1. Top panels: observed photometry with the colours J × ( J − H ) (left) and J × ( J − K s ) (right). Middle: equal-area(29 . ′ < R < ′ ) extraction from the comparison field, wherethe disc contamination is present. Bottom panels: decontaminatedCMDs that suggest a relatively reddened and distant MSTO, redclump, and giant branch typical of old OCs, fitted with the 1 GyrSolar-metallicity Padova isochrone. The shaded polygon corre-sponds to the colour-magnitude filter (Sect. 6). Arrows in thebottom panels show the reddening vector computed for A V = 2. tially on two different premises. The first works with spatialvariations of the star-count density, but does not take intoaccount CMD properties. In the latter, stars in a CMD ex-tracted from an assumed cluster region are subtracted ac-cording to colour and magnitude similarity with the starsof an equal-area comparison field CMD. These methods, to-gether with the one we work with, are based on photomet-ric properties only. Ideally, more robust results on clustermembership determination would be obtained if another in-dependent parameter, such as the proper motion of memberand comparison field stars, is taken into account. However,for proper motions to be useful the target cluster should berelatively close (e.g. Alessi, Moitinho & Dias 2003) and/orto have been observed in widely-apart epochs, preferentiallywith high resolution, such as in the case of the globularcluster NGC 6397 (Richer et al. 2008). Neither condition issatisfied for the present clusters, which are relatively dis-tant (Sect. 5) and have been observed by 2MASS in a singleepoch. S )LK10 Raw (R=5‘)Field stars Same areaClean Figure 5.
Same as Fig. 4 for the region
R < ′ of LK 10. Theequal-area comparison field extraction was taken from the re-gion 29 . ′ < R < ′ . A relatively populous giant clump andabout 3.5 mag of the MS show up, especially in the decontami-nated CMDs, denoting advanced age. The 1 Gyr Solar-metallicityPadova isochrone is applied to the CMDs. Reddening vectors asin Fig. 4. We work with the statistical algorithm introduced byBonatto & Bica (2007a) to deal with the field-star contami-nation in CMDs. The algorithm takes into account simulta-neously star-count density and colour/magnitude similaritybetween cluster and comparison field. It measures the rel-ative number densities of probable field and cluster starsin cubic CMD cells whose axes correspond to the J mag-nitude and the ( J − H ) and ( J − K s ) colours . The algo-rithm: (i) divides the full range of magnitude and colourscovered by the CMD into a 3D grid, (ii) calculates the ex-pected number density of field stars in each cell based onthe number of comparison field stars with similar magni-tude and colours as those in the cell, and (iii) subtracts theexpected number of field stars from each cell. By construc-tion, the algorithm is sensitive to local field-star contami-nation (Bonatto & Bica 2007a). Typical cell dimensions are∆ J = 1 .
0, and ∆( J − H ) = ∆( J − K s ) = 0 .
25, which arelarge enough to allow sufficient star-count statistics in indi- These are the 2MASS colours that provide the maximumvariance among CMD sequences for OCs of different ages (e.g.Bonatto, Bica & Girardi 2004). c (cid:13) , 1–15 nvestigation of 4 old infrared OCs S )FSR1521 Raw (R=4’)Field stars Same areaClean Figure 6.
Same as Fig. 4 for the region
R < ′ of FSR 1521, withthe equal-area comparison field extraction taken from 29 . ′ Same as Fig. 4 for the region R < ′ of FSR 1555. Theequal-area comparison field extraction was taken from the region29 . ′ < R < ′ . The decontaminated CMDs are best fittedwith the 1.5 Gyr Solar-metallicity Padova isochrone. Reddeningvectors as in Fig. 4. Additional statistical analysis is required because of the rela-tively high reddening values (Table 2) affecting the clusters.In Table 3 we present the full statistics of the decontami-nation, discriminated by magnitude bins. Statistically rele-vant parameters are: (i) N σ which, for a given magnitudebin, corresponds to the ratio of the decontaminated numberof stars to the 1 σ Poisson fluctuation of the number of ob-served stars, (ii) σ FS , which is related to the probability thatthe decontaminated stars result from the normal star countfluctuation in the comparison field and, (iii) F S unif , whichmeasures the star-count uniformity of the comparison field.Properties of N σ , σ FS , and F S unif , measured in OCs andfield fluctuations are discussed in Bica, Bonatto & Camargo(2008). Table 3 also provides integrated values of the aboveparameters, which correspond to the full magnitude rangespanned by the CMD of each OC. The spatial regions arethose sampled by the CMDs shown in the top panels ofFigs. 4-7.CMDs of star clusters should have integrated N σ val-ues significantly higher than 1 (Bica, Bonatto & Camargo2008), a condition that is met for the present objects ( N σ =6 . − . c (cid:13) , 1–15 C. Bonatto and E. Bica Table 3. Field-star decontamination statistics∆ J LK 1 ( R < ′ ) - f sub = 97 . 5% LK 10 ( R < ′ ) - f sub = 91 . N obs N cl N σ σ FS F S unif N obs N cl N σ σ FS F S unif (mag) (stars) (stars) (stars) (stars) (stars) (stars)8–9 — — — — — 1 ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . J FSR 1521 ( R < ′ ) - f sub = 97 . 9% FSR 1555 ( R < ′ ) - f sub = 96 . N obs N cl N σ σ FS F S unif N obs N cl N σ σ FS F S unif (mag) (stars) (stars) (stars) (stars) (stars) (stars)8–9 2 ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . ± . J ), we give the number of observed stars ( N obs ) within the spatial region sampled in the CMDsshown in the top panels of Figs. 4 and 5, the respective number of probable member stars ( N cl ) computed by the decontaminationalgorithm, the N σ parameter, the 1 σ Poisson fluctuation ( σ FS ) around the mean, with respect to the star counts measured in the 8sectors of the comparison field, and the field-star uniformity parameter. The statistical significance of N cl is reflected in its ratio with the1 σ Poisson fluctuation of N obs ( N σ ) and with σ FS . The bottom line corresponds to the full magnitude range. The subtraction efficiency( f sub ) is also given. rameter σ FS , which is the 1 σ Poisson fluctuation around themean of the star counts measured in the 8 sectors (correctedfor the different areas of the sectors and cluster extraction).In a spatially uniform comparison field, σ FS is expected tobe very small. Thus, OCs should have the probable num-ber of member stars ( N cl ) higher than ∼ σ FS , to minimisethe probability that N cl arises from fluctuations of a non-uniform comparison field. This condition is fully satisfied bythe present clusters, reaching the level N cl ∼ (5 − σ FS .We also provide in Table 3 the parameter F S unif . For a givenmagnitude bin we first compute the average number of starsover all sectors h N i and the corresponding 1 σ fluctuation σ h N i ; thus, F S unif is defined as F S unif = σ h N i / h N i . Nonuniformities such as heavy differential reddening should re-sult in high values of F S unif . On the other hand, F S unif islow ( . . F S unif = 0 . N cl & σ FS ).Since we usually work with comparison fields largerthan the possible-cluster extractions, the correction for thedifferent spatial areas between field and cluster is expectedto produce a fractional number of probable field stars ( n cellfs ) in some cells. Before the cell-by-cell subtraction, the frac-tional numbers are rounded off to the nearest integer, butlimited to the number of observed stars in each cell n cellsub = NI ( n cellfs ) n cellobs , where NI represents rounding off to thenearest integer). The global effect is quantified by means ofthe difference between the expected number of field stars ineach cell ( n cellfs ) and the actual number of subtracted stars( n cellsub ). Summed over all cells, this quantity provides an es-timate of the total subtraction efficiency of the process, f sub = 100 × X cell n cellsub / X cell n cellfs (%) . Ideally, the best results would be obtained for an efficiency f sub ≈ As an indicator of the algorithm efficiency we can take thedecontaminated stellar surface-density distributions (bot-tom panels of Figs. 2 and 3). The central excesses havebeen significantly enhanced with respect to the raw photom-etry (top panels), while the residual surface-density around c (cid:13) , 1–15 nvestigation of 4 old infrared OCs the centre has been reduced to a minimum level. By de-sign, the decontamination depends essentially on the colour-magnitude distribution of stars located in different spatialregions. The fact that the decontaminated surface-densitypresents a conspicuous excess only at the assumed clusterposition implies significant differences among this region andthe comparison field, both in terms of colour-magnitude andnumber of stars within the corresponding colour-magnitudebins. This meets the expectations of star clusters, whichcan be characterised by a single-stellar population, projectedagainst a Galactic stellar field.The decontaminated CMDs are shown in the bottompanels of Figs. 4 - 7. As expected, essentially all of the disccontamination is removed, leaving stellar sequences typicalof reddened old OCs, with well-developed red clumps anddifferent extents of the MS.As a caveat, we cannot exclude the possibility of differ-ential reddening to account for part of the observed spreadin the CMD distribution of stars. To examine this issue weshow, in the bottom panels of Figs. 4-7, reddening vectorscomputed with the 2MASS ratios (Sect. 4) for a standardvisual absorption A V = 2. Given the absorptions derived forthe clusters (Table 2), the standard value, which representsfrom ≈ 25% to ≈ 50% of the total A V , can be taken asan upper limit to the differential reddening. Together withthe decontaminated CMDs, this experiment suggests thatdifferential reddening in all cases is not large, because thegiant clumps are rather tight, while the MS spread appearsto be dominated by photometric errors. The typical depen-dence of the 2MASS photometric errors on magnitude, forobjects projected along different directions, is discussed inBonatto & Bica (2007b).We conclude that the qualitative and quantitative ex-pectations of the decontamination algorithm have been sat-isfied by the output. In all cases, the decontaminated pho-tometry presents a conspicuous excess, with respect to thesurroundings, in the surface-density distribution (Figs. 2and 3). In addition, field-decontaminated CMDs extractedfrom the spatial regions where the excesses occur (Figs. 4-7),present statistically significant (Table 3) cluster CMDs. The field-decontaminated CMD morphologies derived inSect. 4 can be used to compute the cluster funda-mental parameters. We work with Padova isochrones(Girardi et al. 2002) computed with the 2MASS J , H , and K s filters . The updated isochronesare very similar to the Johnson-Kron-Cousins ones (e.g.Bessel & Brett 1988), with differences of at most 0.01 in( J − H ) (Bonatto, Bica & Girardi 2004). Distinctive fea-tures of the updated isochrone set are centred mostly onthe greatly-improved treatment of the thermally-pulsingasymptotic giant branch (TP-AGB) phase. According toMarigo et al. (2008), the updated isochrones are intended http://stev.oapd.inaf.it/cgi-bin/cmd - Bolometric and colourcorrections were computed for a set of isochrones using the2MASS filter responses. The isochrones were subsequently pro-vided in the Vega Mag system. to preserve the several peculiarities present in the TP-AGBtracks, namely, the cool tails of C-type stars due to the useof proper molecular opacities as convective dredge-up oc-curs along the TP-AGB, the bell-shaped sequences in theHR diagram for stars with hot-bottom burning, the changesof pulsation mode between fundamental and first overtone,the sudden changes of mean mass-loss rates as the surfacechemistry changes from M- to C-type, etc. Because it is im-portant to derive, or at least set constraints on astrophysicalparameters of old star clusters, we adopt as working strategythe search for solutions within a range of ages and metallic-ities. As discussed by, e.g. Friel (1995), OC metallicities ingeneral range from Solar ([ F e/H ] = 0, or Z = 0 . F e/H ] = − . Z = 0 . ≈ / F e/H ] ≈ − . 1. Thus we base the following anal-ysis on this most probable value. To compute Galactocentricdistances, we adopt R ⊙ = 7 . ± . .Historically, different approaches have been used to ex-tract astrophysical parameters from CMDs by means ofisochrone fits. The simplest ones are based on a direct com-parison of a set of isochrones with the CMD morphology,while the more sophisticated include photometric uncer-tainties, binarism, and variations on metallicity. Most ofthese methods are summarised in Naylor & Jeffries (2006),in which a maximum-likelihood CMD fit method is de-scribed.For simplicity, in the present cases fits are made by eye ,with the tight giant clumps as the strongest constraint. Wealso require that, because of the potential presence of bi-naries, the adopted isochrone should be shifted somewhatto the left of the MS fiducial line, i.e. a median line thattakes into account the MS spread, including the photometricuncertainties as well (e.g. Bonatto, Bica & Santos Jr. 2005,and references therein). In the following Section we discusseach cluster individually. The decontaminated CMD morphology of FSR 1521 (Fig. 6)shows a concentration of stars at J ≈ 15, which indicates arelatively populous MSTO of an old cluster detected nearthe 2MASS photometric limit. Besides, a distinctive redclump shows up at J ≈ . J − H ) ≈ . 9. Takentogether, these stellar sequences characterise a distant oldOC. Allowing as well for the photometric uncertainties, ac-ceptable fits to the decontaminated CMD morphology areobtained with the 2 Gyr isochrone, with an uncertainty of ± . F e/H ] = − . 5) metallicities. Agesyounger than about 1.5 Gyr would produce a significantly Other recent studies gave similar results, e.g. R ⊙ = 7 . ± . R ⊙ = 7 . ± . 32 kpc (Eisenhauer et al.2005) and R ⊙ = 7 . ± . 10 kpc (Nishiyama et al. 2006), withdifferent approaches.c (cid:13) , 1–15 C. Bonatto and E. Bica Figure 8. Possible solutions for the age and metallicity ofFSR 1521. Padova isochrones with the ages 1.5 (left panels), 2(middle), and 2.5 Gyr (right) are used. Except for the Solar metal-licity 1.5 Gyr isochrone, the remaining ones provide acceptablefits, either with Solar-metallicity (top panels) or the 1 / poorer description (much fainter than observed) of the redclump, while those older than about 2.5 Gyr, would be pro-hibitively shifted to the right of the MSTO. In any case,both metallicity ranges provide similar fits to the decon-taminated morphology. Since the remaining OCs have ages(Table 2) and CMDs with photometric uncertainties sim-ilar to those of FSR 1521, the above test shows that anymetallicity within 1 / . Z/Z ⊙ . E ( J − H ) =0 . ± . 05, which corresponds to E ( B − V ) = 1 . ± . 2, or A V = 3 . ± . 5, the observed and absolute distance moduli( m − M ) J = 14 . ± . m − M ) O = 13 . ± . 33, re-spectively, and the distance from the Sun d ⊙ = 4 . ± . R ⊙ = 7 . R GC = 7 . ± . ≈ . The best-fit to the CMD of LK 1 was obtained with the1 Gyr isochrone, E ( J − H ) = 0 . ± . 03 and ( m − M ) J =15 . ± . 2. Taking into account fit uncertainties we derivethe age 1 . ± . E ( B − V ) = 2 . ± . A V = 8 . ± . m − M ) O = 13 . ± . d ⊙ = 4 . ± . R GC =7 . ± . ≈ . d ⊙ ≈ The decontaminated CMD morphology of LK 10 (Fig. 5)is somewhat more constrained than that of LK 1. Besidesa tight red clump, it features about a 3 mag MS extent.Fundamental parameters of LK 10 are an age 1 . ± . E ( J − H ) = 0 . ± . 02, which corresponds to E ( B − V ) =2 . ± . A V = 8 . ± . 3, ( m − M ) J = 13 . ± . m − M ) O = 10 . ± . d ⊙ = 1 . ± . R GC =7 . ± . ≈ . d ⊙ ∼ . With the MSTO and red clump detected in the decon-taminated CMD of FSR 1555 (Fig. 7), we derive an age1 . ± . E ( J − H ) = 0 . ± . 02, which correspondsto E ( B − V ) = 1 . ± . A V = 3 . ± . 3, ( m − M ) J =14 . ± . 1, ( m − M ) O = 13 . ± . d ⊙ = 4 . ± . R GC = 7 . ± . ≈ . c (cid:13) , 1–15 nvestigation of 4 old infrared OCs Structural parameters are derived by means of the projectedradial density profiles (RDP) built with the stellar num-ber density around the cluster centre. Usually, star clus-ters have RDPs that follow a well-defined analytical pro-file. Among these are the empirical, single mass, modifiedisothermal sphere of King (1966), the modified isothermalsphere of Wilson (1975), which assumes a pre-defined stel-lar distribution function (and produces more extended en-velopes than King 1966), and the power law with a coreof Elson, Fall & Freeman (1987). These functions are char-acterised by different parameters that are related to clus-ter structure. However, considering the error bars of thepresent RDPs (Fig. 9), we adopt the analytical function σ ( R ) = σ bg + σ / (1 + ( R/R c ) ), where σ bg is the residualbackground density, σ is the central density of stars, and R c is the core radius. This function is similar to that in-troduced by King (1962) to describe the surface brightnessprofiles in the central parts of globular clusters. As discussedin Bonatto & Bica (2008a), RDPs built with depth-limitedphotometry produce structural radii comparable to the in-trinsic (i.e. derived with deep photometry) ones.To minimise noise in the RDPs, we first apply a colour-magnitude filter to the photometry, which excludes starswith colours unlike those of the cluster sequence. Colour-magnitude filters are wide enough to include cluster MS andevolved star colour distributions, as well as the 1 σ photo-metric uncertainties . The colour-magnitude filters for thepresent OCs are shown in the bottom-left panels of Figs. 4-7. However, residual field stars with colours similar to thoseof the cluster are expected to remain inside the colour-magnitude filter. They affect the intrinsic stellar RDP in away that depends on the relative densities of field and clus-ter stars. The contribution of the residual contaminationto the observed RDP is statistically subtracted by meansof the field. As a result, the use of colour-magnitude filtersenhances the contrast of the RDP with respect to the back-ground, especially in crowded fields (e.g. Bonatto & Bica2007a).Oversampling near the centre and undersampling atlarge radii are avoided by using rings of increasing widthwith distance from the cluster centre. A typical set ofring widths is ∆ R = 0 . , , . , and 5 ′ , respectively for0 ′ R < ′ , 1 ′ R < ′ , 4 ′ R < ′ , and R > ′ . Thenumber and width of the rings can be set to produce RDPswith adequate spatial resolution and small 1 σ Poisson errors.The residual background level of each RDP corresponds tothe average number of colour-magnitude filtered stars mea-sured in the field. The R coordinate (and uncertainty) ofeach ring corresponds to the average position and standarddeviation of the stars inside the ring.The colour-magnitude filtered RDPs of the present clus-ters are shown in Fig. 9, where we also show the pro-files produced with the observed (raw) photometry. As ex- Colour-magnitude filter widths should also account for forma-tion or dynamical evolution-related effects, such as enhanced frac-tions of binaries (and other multiple systems) towards the centralparts of clusters, since such systems tend to widen the MS (e.g.Hurley & Tout 1998; Kerber et al. 2002; Bonatto & Bica 2007a;Bonatto, Bica & Santos Jr. 2005). pected, minimisation of the number of non-cluster stars bythe colour-magnitude filter resulted in RDPs with highercontrast with respect to the background. Fits of the King-like profile were performed with a non-linear least-squaresfit routine that uses errors as weights. To minimise degreesof freedom, σ and R c were derived from the RDP fit, while σ bg is measured in the field. The best-fit solutions are shownin Fig. 9, and the fit parameters are given in Table 4. For ab-solute comparison with other clusters, Table 4 also presentsparameters in absolute units, based on the cluster distances(Sect. 5). Because of the 2MASS photometric limit, whichfor the present clusters corresponds to a cutoff for starsbrighter than J ≈ . σ should be taken as a lower limit.Within uncertainties, the adopted King-like functiondescribes well the colour-magnitude filtered RDPs alongthe full radius range, especially for LK 1, FSR 1521, andFSR 1555. The exception is LK 10, which shows a marked ex-cess in the central region. This central cusp in LK 10 suggestsa post-core collapse phase in this ∼ R RDP ) by visu-ally comparing the RDP level (and fluctuations) with thebackground. It corresponds to the distance from the clus-ter centre where RDP and background are statistically in-distinguishable (e.g. Bonatto & Bica 2005, and referencestherein). Thus, most of the cluster stars are contained within R RDP , which should not be mistaken for the tidal radius.Tidal radii are derived from, e.g. the 3-parameter King-profile fit to RDPs (see below), which requires large sur-rounding fields and adequate Poisson errors. For instance,in populous and relatively high Galactic latitude OCs suchas M 67, NGC 188, and NGC 2477, cluster radii are a fac-tor ∼ . − . δ c = 1 + σ /σ bg , which, for the present clusters is rela-tively high (3 . < δ c < . δ c is measured in colour-magnitude-filtered (lower noise) RDPs, it is usually higherthan the visual contrast produced by images (e.g. Fig. 1).Alternatively, we tried to fit the RDPs with the 3-parameter function (based on King 1962) σ ( R ) = σ " p R/R c ) − p R t /R c ) , which includes the tidal radius ( R t ). However, convergenceoccurred only for LK 1, with R t = 18 . ′ ± . ′ . Withsuch radii, the concentration parameter of LK 1 is c =log( R t /R c ) ≈ . c (cid:13) , 1–15 C. Bonatto and E. Bica Table 4. Derived cluster structural parametersCluster σ bg σ R c R RDP R t δ c ′ σ bg σ R c R RDP R t ( ∗ ′− ) ( ∗ ′− ) ( ′ ) ( ′ ) ( ′ ) (pc) ( ∗ pc − ) ( ∗ pc − ) (pc) (pc) (pc)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)LK 1 5 . ± . 03 38 . ± . . ± . 10 6 . ± . ± . ± . . ± . . ± . . ± . 12 7 . ± . ± . ± . 02 9 . ± . . ± . 57 16 . ± . . ± . . ± . . ± . . ± . 23 6 . ± . . ± . 06 15 . ± . . ± . 24 5 . ± . . ± . . ± . . ± . . ± . 31 6 . ± . . ± . 03 37 . ± . . ± . 05 4 . ± . . ± . . ± . . ± . . ± . 06 5 . ± . R c ), cluster ( R RDP ), and tidal ( R t ) radii are given in angular and absolute units. Col. 7: cluster/background densitycontrast parameter ( δ c = 1 + σ /σ bg ), measured in the colour-magnitude filtered RDPs. Col. 8: arcmin to parsec scale. σ ( s t a r s a r c m i n − ) σ ( s t a r s p c − ) LK1 LK10FSR1555FSR1521R RDP R c R RDP R c R c R c R RDP (a) (b)(c) (d) R RDP R t Figure 9. Stellar RDPs built with colour-magnitude filtered(filled circles) and raw (empty) photometry. Solid line: best-fitKing-like profile. Horizontal shaded polygon: background stellarlevel. Shaded regions: 1 σ King fit uncertainty. The core ( R c ) andcluster ( R RDP ) radii are indicated in all cases, while for LK 1, thetidal radius ( R t ) is also shown. Note the central density excess inLK 10 profile. with colour-magnitude-filtered RDPs, which enhances thecluster/background contrast and probes larger cluster ex-tensions. With respect to FSR 1521 and FSR 1555, our val-ues for the core radius are about 1 / R RDP is con-sistent with their R t , given the above relation between bothradii.Compared to the distribution of core radius derived fora sample of relatively nearby OCs by Piskunov et al. (2007),the present OCs, especially FSR 1521, occupy the small- R c tail. Besides, for a relation between tidal and cluster radiusas ∼ × R RDP , LK 10, FSR 1521, and FSR 155 are locatedaround the median value, while LK 1 occupies the large- R t tail. Since we detect ≈ (cid:0) φ ( m ) = dNdm (cid:1) and compute themass stored in stars. We work with colour-magnitude fil-tered photometry, the 3 2MASS bands separately, andthe mass-luminosity relation obtained from the corre-sponding Padova isochrone and distance from the Sun(Sect. 5). Further details on MF construction are given inBica, Bonatto & Blumberg (2006). The effective MS stellarmass range is 1 . m ( M ⊙ ) . φ ( m ) ∝ m − (1+ χ ) , with the slope χ = 2 . ± . 4. Withinuncertainties, this slope agrees with χ = 1 . ± . 25 derivedby Le Duigou & Kn¨odlseder (2002). Both values are steeperthan the χ = 1 . 35 of Salpeter (1955) initial mass function(IMF).The number of observed MS and evolved stars in LK 10(for R R RDP ) is derived by counting the stars (in thebackground-subtracted colour-magnitude filtered photome-try) that are present in the mag ranges, 13 . < J < . J < . n MS = 959 ± 20 and n evol = 25 ± 8, MS and evolved stars,respectively; the corresponding mass values are m MS =1311 ± M ⊙ and m evol = 46 ± M ⊙ . The evolved starmass corresponds to n evol multiplied by the stellar mass atthe TO, m TO = 1 . M ⊙ . Thus, the observed stellar mass ofLK 10 is m obs ≈ M ⊙ , which agrees with the equivalentvalue estimated by Le Duigou & Kn¨odlseder (2002).Finally, we estimate the total stellar mass by extrapo-lating the observed MF down to the H-burning mass limit(0 . M ⊙ ). We follow the universal IMF of Kroupa (2001),which assumes increasing flattening towards low-mass stars.This IMF is described by the slopes χ = 0 . ± . . m ( M ⊙ ) . χ = 1 . ± . . m ( M ⊙ ) . 0. We obtain m extr = 4420 ± M ⊙ . Thus,the stellar mass of LK 10 can be put in the range 1360 − M ⊙ . Again, within uncertainties, the upper value isconsistent with that estimated by Le Duigou & Kn¨odlseder(2002).Because of the limited MS range, only estimates of the c (cid:13) , 1–15 nvestigation of 4 old infrared OCs observed cluster mass are made for the remaining objects,by means of the age solutions given in Sect. 5. In all cases weconsider the region within R R RDP (Table 4). For LK 1we derive m MS = 108 ± M ⊙ and m evol = 270 ± M ⊙ ,which leads to the total observed mass m obs ≈ M ⊙ . Thisvalue corresponds to about 1 / m MS = 123 ± M ⊙ , m evol = 139 ± M ⊙ , and m obs ≈ M ⊙ . Since FSR 1555 and LK 1 are approximately at thesame distance from the Sun, the latter appears to be some-what more massive than the former.With the more probable solution (2 Gyr) for FSR 1521,the total number of stars present in R . . ′ CMD is n tot ≈ ≈ 18 are in the red clump and ≈ ≈ . M ⊙ , theobserved mass of FSR 1521 is m obs ≈ M ⊙ , of which ≈ M ⊙ correspond to the evolved stars, and ≈ M ⊙ arestored in the red clump. Interestingly, while the observedmass of FSR 1521 is similar to that of FSR 1555, the evolvedmass is somewhat lower. Since both OCs are at comparabledistances, this difference is consistent with the relative ages.The fact that LK 1, FSR 1521, and FSR 155 present sim-ilar observed masses as LK 10, suggest that they all mightbe as massive as the latter OC.Given the accuracy of the isochrone best-fit for old OCsprovided by near-infrared decontaminated CMDs, e.g. Fig. 8and related discussions, the observed masses are expected tobe representative. Of course, deeper near-infrared photom-etry coupled to similar methods as the present one, and/orincluding proper motion filtering, would produce more con-strained results. With the analyses of the preceding sections we have gath-ered important clues to establish that the objects dealt within this paper are Gyr-class OCs, or older. We also derivedrepresentative, i.e. constrained by means of the isochrone fit(Sect. 5), fundamental and structural parameters, most ofwhich have been derived for the first time. We use these pa-rameters to put the clusters into perspective, by comparingsome of their properties with those of a set of well-studiedOCs.As reference we take the nearby OCs with ages in therange 70 − − M ⊙ studied by Bonatto & Bica (2005), together with theyoung OCs NGC 6611 (Bonatto, Santos Jr. & Bica 2006)and NGC 4755 (Bonatto et al. 2006b). The reference clus-ters are distinguished according to total mass (higher orlower than 1 000 M ⊙ ). Bonatto & Bica (2005) discuss pa-rameter correlations in the reference sample.As shown in panel (a) of Fig. 10, the core and clus-ter radii of the reference OCs are related by R RDP =(8 . ± . × R (1 . ± . , which suggests a similar scalingfor both kinds of radii, at least for the radii ranges 0 . . R c (pc) . . . R RDP (pc) . 15. LK 1, LK 10,FSR 1521, and FSR 1555 fit tightly in the relation. Theyalso appear to follow the trend of increasing cluster radiiwith Galactocentric distance (panel b). This kind of de-pendence was previously suggested by, e.g. Lyng˚a (1982).Part of this relation may be primordial, in the sense that Age (Myr)0.11 R c o r e ( p c ) Age (Myr)110 R RD P ( p c ) LK1LK10FSR1555FSR1521 GC (kpc)110 R RD P ( p c ) core (pc)110 R RD P ( p c ) (a) (b)(c) (d) Figure 10. Relations involving OC structural and fundamen-tal parameters. Circles: reference OCs. Dotted circles: massive( > M ⊙ ) OCs. We caution that LK 10 may be a post-corecollapse open cluster. In this figure we adopted the core radiusderived from the King-like fit. the higher molecular gas density in central Galactic regionsmay have produced clusters with small radii, as suggestedby van den Bergh, Morbey & Pazder (1991) to explain theincrease of globular cluster radii with Galactocentric dis-tance. After formation, mass loss associated with stellar anddynamical evolution (such as mass segregation and evapo-ration), together with tidal interactions with the Galacticpotential and giant molecular clouds, also contribute to thedepletion of star clusters, especially the low-mass and cen-trally located ones (Sect. 1). A similar dependence on Galac-tocentric distance for R c is implied by the data shown inpanel (a).In panels (c) and (d) of Fig. 10 we compare the presentlyderived cluster and core radii with those of the refer-ence sample in terms of age. LK 1, LK 10, FSR 1521, andFSR 1555 have core and cluster radii similar to those mea-sured in the reference OCs of equivalent age.We show in Fig. 11 the spatial distribution of LK 1,LK 10, FSR 1521, and FSR 1555, as they lay in the Galacticplane. The spiral arm structure of the Milky Way is basedon Momany et al. (2006) and Drimmel & Spergel (2001), asderived from HII regions, and molecular clouds (e.g. Russeil2003). The Galactic bar is shown with an orientation of 14 ◦ and 6 kpc in total length (Freudenreich 1998; Vall´ee 2005).The present OCs are compared to the spatial distributionof the OCs with known age given in the WEBDA database.For comparison purposes we consider two age groups, clus- c (cid:13) , 1–15 C. Bonatto and E. Bica −12 −8 −4 0 4 8 12y GC (kpc)−16−12−8−4048 x G C ( k p c ) τ < 1.0Gyr τ > 1.0GyrLK1LK10FSR1555FSR1521SunOuter ArmPerseusCarinaSagittariusCrux−Scutum NormaBar Orion−Cygnus Figure 11. Spatial distribution of the present star clusters com-pared to the WEBDA OCs with ages younger (gray circles)and older than 1 Gyr (black dots). Clusters are overplotted ona schematic projection of the Galaxy, as seen from the Northpole, with 7.2 kpc as the Sun’s distance to the Galactic centre.Main structures are identified. ters younger and older than 1 Gyr (the old OCs listed inTable 1 are merged into this group). As expected, old OCsare found preferentially outside the Solar circle, while theinner Galaxy contains few OCs so far detected. Besides, be-cause of the presence of bright stars, young OCs can bedetected farther than the old ones, especially towards thecentral region. As discussed in Bonatto et al. (2006a), cen-tral directions farther than ≈ ≈ Accurate fundamental and structural parameters of old starclusters are important for reasons that range from the com-pleteness of the open cluster parameter space to the deter-mination of the efficiency of cluster dissolution mechanisms.Taken together, the results of the present paper, and thoseof previous works by our group (Table 1), add to a totalof 18 old (age & . pa- In the context discussed in Sect. 5, with uncertainties propa-gated from the best-fitting 2MASS isochrones. rameters derived, for the first time for most of them. Newfindings and the age-determination within a reasonable con-fidence level are important as well to improve the statisticalcoverage of the open cluster databases, especially for the def-inition of the old-age tail of the open cluster age-distributionfunction.In this paper we use colour-magnitude diagrams andradial density profiles to derive fundamental and structuralparameters of the infrared open clusters LK 1 and LK 10,as well as of the star cluster candidates FSR 1521, andFSR 1555. Our approach is essentially based on field-stardecontaminated 2MASS photometry, which enhances clus-ter CMD evolutionary sequences, and produces more con-strained parameters.We present consistent evidence, in the form of CMDmorphology, statistical tests, structural parameters, andcomparison with nearby OCs, that the objects are Gyr-classopen clusters. With absorptions in the range 3 . A V . 9, the objects are well suited for the 2MASS photometry.They are located at d ⊙ ≈ d ⊙ ≈ . ≈ . χ = 2 . ± . 4) somewhat steeperthan Salpeter’s ( χ = 1 . m ( M ⊙ ) m ( M ⊙ ) ACKNOWLEDGEMENTS The anonymous referee is acknowledged for a thorough read-ing of the original manuscript and for valuable suggestionsthat improved the paper. This publication makes use of dataproducts from the Two Micron All Sky Survey, which isa joint project of the University of Massachusetts and theInfrared Processing and Analysis Centre/California Insti-tute of Technology, funded by the National Aeronautics andSpace Administration and the National Science Foundation.This research has made use of the WEBDA database, op-erated at the Institute for Astronomy of the University ofVienna. We acknowledge support from the Brazilian Insti-tution CNPq. c (cid:13) , 1–15 nvestigation of 4 old infrared OCs REFERENCES Alessi, B.S., Moitinho, A. & Dias, W. S. 2003, A&A, 410,565Baumgardt, H. & Makino, J. 2003, MNRAS, 340, 227van den Bergh, S. 1957, ApJ, 125, 445van den Bergh, S., Morbey, C. & Pazder, J. 1991, ApJ, 375,594Bessel, M.S. & Brett, J.M. 1988, PASP, 100, 1134Bica, E., Bonatto, C. & Dutra, C. 2003, A&A, 405, 991Bica, E., Bonatto, C., Barbuy, B. & Ortolani, S. 2006,A&A, 450, 105Bica, E., Bonatto, C. & Blumberg, R. 2006, A&A, 460, 83Bica, E., Bonatto, C., Ortolani, S. & Barbuy, B. 2007,A&A, 472, 483Bica, E. & Bonatto, C. 2008, MNRAS, 384, 1733Bica, E., Bonatto, C. & Camargo, D. 2008, MNRAS, 385,349Bonatto, C., Bica, E. & Girardi, L. 2004, A&A, 415, 571Bonatto, C., Bica, E. & Santos Jr., J.F.C. 2005, A&A, 433,917.Bonatto, C. & Bica, E. 2005, A&A, 437, 483Bonatto, C., Santos Jr., J.F.C. & Bica, E. 2006, A&A, 445,567Bonatto, C., Kerber, L.O., Bica, E. & Santiago, B.X. 2006a,A&A, 446, 121Bonatto, C., Bica, E., Ortolani, S. & Barbuy, B. 2006b,A&A, 453, 121Bonatto, C. & Bica, E. 2006, A&A, 455, 931Bonatto, C. & Bica, E. 2007a, MNRAS, 377, 1301Bonatto, C., Bica, E., Ortolani, S. & Barbuy, B. 2007, MN-RAS, 381, L45Bonatto, C. & Bica, E. 2007b, A&A, 473, 445Bonatto, C. & Bica, E. 2008a, A&A, 477, 829Bonatto, C. & Bica, E. 2008b, A&A, 479, 741Bonatto, C. & Bica, E. 2008c, A&A, 485, 81Bonatto, C., Bica, E. & Santos Jr., J.F.C. 2008, MNRAS,386, 324Bonatto, C. & Bica, E. 2008d, A&A, accepted (astro-ph/0809.2492)Cardelli, J.A., Clayton, G.C. & Mathis, J.S. 1989, ApJ,345, 245Drimmel, R., & Spergel, D.N. 2001, ApJ, 556, 181Le Duigou, J.M. & Kn¨odlseder, J. 2002, A&A, 392, 869Dutra, C.M. & Bica, E. 2001, A&A, 376, 434Dutra, C.M., Santiago, B.X. & Bica, E. 2002, A&A, 383,219Eisenhauer, F., Sch¨odel, R, Genzel, R., Ott, T., Tecza, M.,Abuter, R., Eckart, A. & Alexander, T. 2003, ApJ, 597,L121Eisenhauer, F., Genzel, R., Alexander, T. et al. 2005, ApJ,628, 246Elson, R.A.W., Fall, S.M. & Freeman, K.C. 1987, ApJ, 323,54Freudenreich, H.T. 1998, ApJ, 492, 495Friel, E.D. 1995, ARA&A 1995, 33, 381Froebrich, D., Scholz, A. & Raftery, C.L. 2007, MNRAS,374, 399Froebrich, D., Meusinger, H. & Scholz, A. 2007, MNRAS,377, L54Froebrich, D., Meusinger, H. & Davis, C.J. 2008, MNRAS,383, L45 Girardi, L., Bertelli, G., Bressan, A., et al. 2002, A&A, 391,195Goodwin, S.P. & Bastian, N. 2006, MNRAS, 373, 752von Hoerner, S. 1958, Astrophys., 44, 221Hurley, J. & Tout, A.A. 1998, MNRAS, 300, 977Janes, K.A. & Phelps, R.L. 1994, AJ, 108, 1773Kerber, L.O., Santiago, B.X., Castro, R. & Valls-Gabaud,D. 2002, A&A, 390, 121Khalisi, E., Amaro-Seoane, P. & Spurzem, R. 2007, MN-RAS, 374, 703King, I. 1962, AJ, 67, 471King, I. 1966, AJ, 71, 64Kroupa, P. 2001, MNRAS, 322, 231Lamers, H.J.G.L.M. & Gieles, M. 2006, A&AL, 455, 17Lyng˚a, G. 1982, A&A, 109, 213Marigo, P., Girardi, L., Bressan, A., Groenewegen, M.A.T.,Silva, L. & Granato, G.L. 2008, A&A, 482, 883Massey, P., Johnson, K.E. & DeGioia-Eastwood, K. 1995,ApJ, 454, 151Mercer, E.P., Clemens, D.P., Meade, M.R., Babler, B.L.,Indebetouw, R., Whitney, B.A., Watson, C., Wolfire,M.G. et al. 2005, ApPJ, 6335, 560Momany, Y., Zaggia, S., Gilmore, G., Piotto, G., Carraro,G., Bedin, L.R. & de Angeli, F. 2006, A&A, 451, 515Momany, Y., Bica, E., Barbuy, B., Ortolani, S. & Bonatto,C. MNRAS, 2008, submittedNaylor, T. & Jeffries, R.D. 2006, MNRAS, 373, 1251Nishiyama, S., Nagata, T., Sato, S. et al. 2006, ApJ, 647,1093Oort, J.H. 1958, in Ricerche Astronomiche , 5, 415, SpecolaVaticana, Proc. of a Conference at Vatican Observatory,Castel Gandolfo, May 20-28, 1957, ed. D.J.K. O’ConnellOrtolani, S., Bica, E., Barbuy, B. & Zocalli, M. 2005, A&A,439, 1135Ortolani, S., Bica, E. & Barbuy, B. 2005, A&A, 437, 531Pavani, D.N. & Bica, E. 2007, MNRAS, 468, 139Piskunov, A.E., Kharchenko, N.V., R¨oser, S., Schilbach, E.& Scholz, R.-D. 2007, A&A, 445, 545Portegies Zwart, S.F., Makino, J., McMillan, S.L.W. &Hut, P. 2002, ApJ, 565, 265Richer, H.B., Dotter, A., Hurley, J., Anderson, J., King, I.,Davis, S., Fahlman, G.G. Hansen, B.M.S. et al. 2008, AJ,135, 2141Russeil, D. 2003, A&A, 397, 133Salaris, M., Weiss, A. & Percival, S.M. 2004,A&A, 422, 217Salpeter, E. 1955, ApJ, 121, 161Skrutskie, M., Schneider, S.E., Stiening, R., et al. 1997,in The Impact of Large Scale Near-IR Sky Surveys , ed.Garzon et al., Kluwer (Netherlands), 210, 187Spitzer, L. 1958, ApJ 127, 17Trager, S.C., King, I.R. & Djorgovski, S. 1995, AJ, 109,218Upgren, A.R., Mesrobian, W.S. & Kerridge, S.J. 1972, AJ,77, 74Vall´ee, J.P. 2005, AJ, 130, 569Wilson, C.P. 1975, AJ, 80, 175 c (cid:13)000