Evidence of the inhomogeneity of the stellar population in the differentially reddened globular cluster NGC 3201
aa r X i v : . [ a s t r o - ph . GA ] A p r Astronomy&Astrophysicsmanuscript no. 13749 c (cid:13)
ESO 2018September 27, 2018 L etter to the E ditor Evidence of the inhomogeneity of the stellar population in thedifferentially reddened globular cluster NGC 3201. ⋆ V. Kravtsov , , G. Alca´ıno , G. Marconi , and F. Alvarado Instituto de Astronom´ıa, Universidad Cat´olica del Norte, Avenida Angamos 0610, Antofagasta, Chilee-mail: [email protected] Sternberg Astronomical Institute, University Avenue 13, 119899 Moscow, Russia Isaac Newton Institute of Chile, Ministerio de Educaci´on de Chile, Casilla 8-9, Correo 9, Santiago, Chilee-mail: [email protected], [email protected] ESO - European Southern Observatory, Alonso de Cordova 3107, Vitacura, Santiago, Chilee-mail: [email protected]
Received xxxxx / Accepted xxxxx
ABSTRACT
Aims.
We report on evidence of the inhomogeneity (multiplicity) of the stellar population in the Galactic globular cluster (GC) NGC3201, which is irregularly reddened across its face.
Methods.
We carried out a more detailed and careful analysis of our recently published new multi-color photometry in a wide field ofthe cluster with particular emphasis on the U band. Results.
Using the photometric data corrected for di ff erential reddening, we found for the first time two key signs of the inhomogeneityin the cluster’s stellar population and of its radial variation in the GC. These are (1) an obvious trend in the color-position diagram,based on the ( U − B ) color-index, of red giant branch (RGB) stars, which shows that the farther from the cluster’s center, the blueron average the ( U − B ) color of the stars is; and (2) the dependence of the radial distribution of sub-giant branch (SGB) stars in thecluster on their U magnitude, where brighter stars are less centrally concentrated than their fainter counterparts at a confidence levelvarying between 99.2% and 99.9% depending on the color-index used to select the stars. The same e ff ects were recently found by usin the GC NGC 1261. However, contrary to NGC 1261, we are not able to unambiguously suggest which of the sub-populations ofSGB / RGB stars can be the progenitor of blue and red horizontal branch stars of the cluster. Apart from M4, NGC 3201 is anotherGC very probably with an inhomogeneous stellar population, which has essentially lower mass than the most massive Galactic GCswhere multiple stellar populations were unambiguously detected for the first time.
Key words. globular clusters: general – globular clusters: individual: NGC 3201
1. Introduction
The southern Galactic globular cluster (GC) NGC 3201, knownnot only by its peculiar kinematic characteristics but also by ir-regular di ff erential reddening across its face, was the subject ofour recent study (Kravtsov et al. 2009) based on a new multi-color photometry in a 14 ′ x14 ′ field of the GC. In that study,where we primarily dealt with some aspects of the properties andcharacteristics of the cluster stellar population, we also alloweda possible spread in the population, but did not examine it. Ina later more detailed analysis of the same data, we were able tofind not only apparent manifestations, but also stronger and moreobjective evidence of the inhomogeneity in the stellar populationand of its radial variation in the cluster. The present letter reportson these findings in detail. The obtained results contribute moreto our past (Alca´ıno et al. 1999) and recent (Kravtsov et al. 2010)studies of the inhomogeneity (multiplicity) of the stellar popu-lations in the populous Large Magellanic Cloud (LMC) clusterNGC 1978 and Galactic GC NGC 1261, respectively and to therapidly growing body of photometric and spectroscopic evidence Send o ff print requests to : V. Kravtsov ⋆ Based on observations with the 1.3 m Warsaw telescope at LasCampanas Observatory about multiple stellar populations in both Magellanic Clouds starclusters (e.g., Mackey et al. 2008; Milone et al. 2009a; and refer-ences therein) and Galactic GCs (some relevant publications arereferred to elsewhere in the paper).
2. The used photometric data and the correctionsapplied to them
In the present study, we are relying on our recent multi-colorphotometry in
U BVI of more than 12 000 stars in a 14 ′ x14 ′ clus-ter field, reaching below the turno ff point in all passbands. Fora description of the photometric data used here (and publiclyavailable in electronic form) and of the typical r.m.s errors ineach passband, see Kravtsov et al. (2009). In the same paper, wealso described in detail our study of the variation of reddeningin the observed cluster field and listed the estimated correctionscorresponding to areas with conditionally defined grades of red-dening. These corrections were estimated separately for the U and V magnitudes, as well as for the ( B − V ), ( V − I ), and ( B − I )colors, and typical uncertainties were quoted. As for the correc-tions for the ( U − B ), they were supplementary calculated in thepresent study from those originally obtained for the ( B − V ) colorby applying the following relation between the color-excesses in Please give a shorter version with: \authorrunning and / or \titilerunning prior to \maketitle both colors: E U − B = E B − V . In the analysis presented below,we used photometric data and the cluster’s CMDs corrected fordi ff erential reddening.In addition to its variable reddening, the cluster field is alsopopulated by the large number of field stars, which are fairly nu-merous in the region of the lower red giant branch (RGB) andsub-giant branch (SGB). A decontamination procedure was ap-plied to the photometry used. For more detail, interested readersare referred to our original paper on photometry of NGC 3201.
3. The sub-giant branch in the U magnitude In two rows of panels of Fig. 1, we demonstrate the apparentradial variation of the U -level of the SGB of NGC 3201 in the U -based CMDs with di ff erent color-indexes. The CMDs are cor-rected for di ff erential reddening and decontaminated of the ma-jority of field stars in the region of the lower RGB and of theSGB. The CMDs shown in the upper and lower panels corre-spond to the inner (1 . ′ < R < . ′
70) and the outer ( R > . ′ U -level and perhaps the slope of the SGB vary with the distancefrom the center of NGC 3201, the SGB being systematicallybrighter (and perhaps steeper) in the outer part of the cluster Thedashed horizontal line is drawn at the same U magnitude in bothrows of panels as a reference line to facilitate a comparison ofthe SGB level in the CMDs of the two parts in the cluster.The qualitatively demonstrated radial variations of the dis-tribution of stars on the SGB in the U -based CMDs are sup-ported by more objective indicators based on quantitative esti-mates. They were obtained in the following way. We first iso-lated a sample of stars most probably belonging to the SGB. The U -based cluster CMD with the ( B − I ) color-index was initiallyused to this goal. This color-index provides the largest separationbetween the turno ff (TO) point and the lower RGB, as comparedwith other available color-indices corrected for di ff erential red-dening. The central part of the cluster was not excluded from theanalysis, thanks to only a weak crowding e ff ect in it. Given aslight apparent slope of the SGB in the U -( B − I ) CMD, it wastaken into account in isolating SGB stars. They were selected inthe color range ∆ ( B − I ) = . < B − I < .
00) and themagnitude range ∆ U = U = -0.194( B − I ) + U = -0.194( B − I ) + ∆ U = ∆ U = ∆ U = B − V ) and ( V − I ), in their ranges ∆ ( B − V ) = . < B − V < .
87) and ∆ ( V − I ) = . < V − I < . ∆ U = U -( B − I ) CMD, with the same U mag-nitudes defining the boundaries of these parts at the extremesof each color range as was accepted for the SGB in the threecolor-indices. Note that although the color-indices under consid-eration are not independent (i.e., each of them is a combinationof two other color-indices), the corrections for di ff erential red-dening were originally made independently in each of them. Thenumber of stars falling in the fainter, intermediate, and brighterboxes are 188, 130, and 117 respectively in the U -( B − V ) CMD. Fig. 1.
The U -based CMDs with di ff erent color-indexes, cor-rected for di ff erential reddening and (partially) rid of field stars,demonstrating the systematic variation of the position of theSGB in the U with radial distance from the center of NGC 3201.Upper and lower panels show the CMDs of inner (1 . ′ < R < . ′
70) and outer ( R > . ′
40) regions of the cluster, respectively.The dashed horizontal line is given as a reference to facilitatinga comparison of the SGB in the CMDs.In turn, the boxes of the same sequence of the brightness levelsin the U -( V − I ) CMD contain 165, 125, and 114 stars, respec-tively. For comparison purposes, we also isolated SGB stars inthe V based CMD, namely V -( V − I ). The SGB is apparentlyslightly narrower in the V magnitude, i.e. ∆ V = ∆ U = U -( V − I ) CMD. The accepted upper and lower borders ofthe SGB in the V -( V − I ) were defined by two envelope lines, V = -1.923( V − I ) + V = -1.923( V − I ) + V magnitude into account, we selectedsub-samples of the brightest and faintest SGB stars in slightly re-duced magnitude ranges than in the U magnitude, namely ∆ V = ∆ V = U based CMDs, one can see that the faintest sub-giantsdenoted by (red) filled triangles are more centrally concentratedthan their counterparts of intermediate brightness in the U (blackfilled circles), and obviously yet more centrally concentratedthan the brightest SGB stars (blue filled squares), irrespectiveof the color-index used for the selection of the SGB stars. Thisapparent di ff erence between the distributions is supported by aquantitative estimate based on a Kolmogorov-Smirnov test: thedi ff erence between the radial distributions of the brightest andfaintest SGB stars is statistically significant at 99.8%, 99.2%,and 99.9% confidence levels for the selection made in the U -( B − I ), U -( B − V ), and U -( V − I ) CMDs, respectively. However,the radial distributions of the sub-samples of SGB stars with dif-ferent brightness in the V magnitude are apparently in contrastto those di ff ering by the U magnitude. In particular, the brightestSGB stars in the V passband are even somewhat more centrally lease give a shorter version with: \authorrunning and / or \titilerunning prior to \maketitle Fig. 2.
Upper panels show three sub-samples of the SGB stars isolated in three magnitude ranges: (1) in the U -based CMDs of NGC3201 with di ff erent color-indices, and additionally (2) in the V -( V − I ) CMD, see text for details. All the CMDs are corrected fordi ff erential reddening. Lower panels: a comparison of the cumulative radial distributions of the three sub-samples of SGB stars; redfilled triangles, black filled circles, and blue filled squares denote the sub-samples of the SGB stars with progressively increasingbrightness in the given passbands. In all the panels, the three sub-samples are shown with symbols of the same color.concentrated in the central part of NGC 3201 than their faintestcounterparts. Moreover, the radial distributions of the two sub-samples of SGB stars are di ff erent at 92.1% confidence level.Such a di ff erence is, strictly speaking, statistically insignificant,because the estimated confidence level is less than 95.0%.The above result may imply the following. Even if the de-pendence of the radial distribution of SGB stars on their bright-ness in the V magnitude (at a given color) really exists and hasthe same meaning as the dependence in the U magnitude, it ispresumably less obvious, so that the errors of the applied correc-tions for di ff erential reddening combined with the errors of thephotometry itself are able to reduce or even distort the depen-dence.The high level of statistical significance of the dependence ofthe SGB stars’ radial distribution in NGC 3201 on their U mag-nitude is strong evidence of the inhomogeneity (”multiplicity”)of the cluster stellar population. This high level was also foundby us in another Galactic GC, NGC 1261. Here we note againthat the inhomogeneity of the population of SGB stars resultingfrom the revealed dependence in both GCs resembles a combina-tion of two e ff ects: (1) the split of the SGB into two componentsrevealed by Milone et al. (2008) in the GC NGC 1851 and (2) thedi ff erent radial distribution of stars belonging to the brighter andfainter components of the SGB, found by Zoccali et al. 2009) inthe same GC, i.e. the brighter SGB stars were found to be lesscentrally concentrated and to extend to much larger radial dis-tances in NGC 1851. Note however that Milone et al. (2009b)do not support this finding.As for the revealed di ff erences regarding SGB stars in NGC3201, one cannot draw any definite conclusion about whether ornot they are of a discrete or a continuous character.
4. The red giant branch and its color-positiondiagram
The obtained evidence of the probable inhomogeneity of SGBstars in NGC 3201 is reinforced by the inhomogeneity of RGBstars and its dependence on the radial distance in the cluster.For our analysis of the RGB of NGC 3201, we used a ready-made sample of the most probable RGB stars. Their selectionwas done previously and is described in detail in Kravtsov etal. 2009. Based on this sample we found as in the RGB of NGC1261 a systematically di ff erent location in the U -( U − B ) diagram of RGB stars situated at di ff erent radial distances from the clus-ter center, where the portion of stars bluer in the ( U − B ) system-atically increased towards the cluster outskirts. In Fig. 3 we com-pare in the U -( U − B ) diagram the position of two sub-samplesof RGB stars belonging to two areas at di ff erent mean radialdistances from the cluster center, namely in the radial ranges1 . ′ < R < . ′
70 and 3 . ′ < R < . ′
3. It is evident that RGB starsin the outer region (blue dots) are systematically bluer than inthe inner one.To secure more convincing evidence and more detailed be-havior of the suggested dependence of the ( U − B ) color of RGBstars on their radial distance, R, from the center of NGC 3201,we obtained the so-called color-position diagram (CPD) of theRGB. Note that this diagram was used for the first time in ourstudy of the stellar population in the LMC populous star clusterNGC 1978 (Alca´ıno et al. 1999). First of all, we linearized theRGB in the U -( U − B ) plane by relying on the ready-made sam-ple of the most probable RGB stars. Since the brightest part ofthe RGB, formed by a dozen of stars with U < .
3, is quasi-horizontal in the U -based CMDs, we rejected these stars. We fit-ted the mean locus of the RGB with a polynomial and subtractedfor each star the color of the mean locus at its luminosity levelfrom the star’s color-index. We left nearly all stars of the initialsample. Only a handful of stars with deviations, δ ( U − B ), fromthe mean locus exceeding ± U magnitude were rejected. Note that the numberof rejected stars is so small that this is rather a formal procedurewith negligible impact. The linearized RGB is shown in the U - δ ( U − B ) plane in the upper panel of Fig. 4. With the purposeexplained below, the linearized RGB was arbitrarily divided bythree magnitude ranges (marked by the dashed lines) with com-parable samples of stars, namely: 189 faint stars with U > . . < U < .
8; 113brighter stars with U < .
8. The final step is plotting the de-pendence between the deviations δ ( U − B ) of the RGB stars andtheir radial distances from the center of NGC 3201.The obtained CPD shown in the lower panel of Fig. 4 isnot only another representation of the demonstrated photomet-ric inhomogeneity of RGB stars in NGC 3201, but also a depen-dence containing additional information. It is now evident be-yond doubt that the ( U − B ) color does get bluer, i.e. the δ ( U − B )becomes more negative with increasing radial distance from thecluster center. It is also obvious that all the three sub-samples of Please give a shorter version with: \authorrunning and / or \titilerunning prior to \maketitle Fig. 3.
Comparison of the location in the U -( U − B ) CMD ofRGB stars from the inner (red dots, 1 . ′ < R < . ′
70) and outer(blue dots, 3 . ′ < R < . ′
3) regions of NGC 3201.RGB stars follow the same trend. Note yet that the CPD assumesthe trend apparently reveals itself beginning at R ≈ − . ′ − . ′
70) and there is no obvious trend within thisradial distance, where the bulk of stars have positive deviations δ ( U − B ). Interestingly enough, this radius is virtually equal tothe cluster half-mass radius, R h = . ′
68 (Harris 1996). In thisconnection it should be also noted that the di ff erence in radialdistribution among SGB stars arises approximately at the sameradius.Concerning the physical reasons that can probably be re-sponsible for the revealed inhomogeneity, we refer to recentlyobtained results by Marino et al. (2008) on stellar population inGC M4. Based on spectroscopy of a sample of 105 stars in thisGC they found a dichotomy in Na abundance and argued that itmust be associated with a CN bimodality. From photometry ofthe same stars Marino et al. (2008) showed that the CN-weak redgiants with a lower content of Na are on average systematicallybluer, by ∆ ( U − B ) = .
17 in the U -( U − B ) CMD, than their CN-strong counterparts with a higher content of Na. This photomet-ric e ff ect is very similar to that revealed by us in NGC 3201 andin NGC 1261. From the CPD we evaluate the mean separation inthe ( U − B ) color between ”blue” and ”red” RGB stars, assum-ing this bi-modality in their distribution in the ( U − B ) color. It isapparently around ∆ ( U − B ) ∼ .
12, which is many times largerthan the possible error in the ( U − B ) color (a few hundredth ofmagnitudes in the worse case) caused by the uncertainty in localfluctuations of reddening in the cluster field under consideration.It agrees well with our estimate made for NGC 1261 and is onthe same order of magnitude as the above-mentioned di ff erencebetween the ”redder” and ”bluer” giants in M4. A tentative firstexplanation is to accept the same reason behind the segregationof RGB stars in the color (U-B) in NGC 3201 as in M4.In NGC 3201, other than in NGC 1261, we are not able tosuggest an unambiguous association between the discussed sub-populations of SGB / RGB stars and those belonging to the blueand red horizontal branch (BHB, RHB).
5. Conclusions
Based on a more detailed analysis of a new multi-color pho-tometry in an extended field of the di ff erentially reddened GCNGC 3201, we found the following signs of the inhomogene-ity (multiplicity) of the cluster’s stellar population. First, thereis an obvious dependence of the radial distribution of SGB starsin the cluster on their U magnitude: brighter stars are less cen-trally concentrated than their fainter counterparts at a confidence Fig. 4.
The trend of the ( U − B ) color of RGB stars with theirradial distance from the center of NGC 3201. Upper panel: thelinearized RGB for the color ( U − B ) corrected for di ff erentialreddening. A dozen of the brightest stars of the RGB ( U < . ff erent symbols. The lower panel shows the color-position diagram of the same RGB stars.level fluctuating above 99.2% in relation to the color-index ofCMD relied on to isolate SGB stars. Second, RGB stars exhibita systematically di ff erent location in both the U -( U − B ) CMDand the R - δ ( U − B ) color-position diagram at di ff erent radial dis-tances from the cluster center: the proportion of stars bluer in the( U − B ) increases towards the cluster outskirts. We note (1) thesame kind of photometric inhomogeneity of RGB and SGB starsin NGC 3201 and in another Galactic GC, NGC 1261 (Kravtsovet al. 2010), and also (2) a very similar radial trend in both GCs.Finally, it is worth mentioning that NGC 3201 after M4, isthe second non-massive Galactic GC that very probably has aninhomogeneous stellar population. Acknowledgements.
We thank the anonymous referee for useful comments thatimproved the manuscript.
References
Alca´ıno, G., Liller, W., Alvarado, F., et al. 1999, A&AS, 135, 103Harris, W. E. 1996, AJ, 112, 1487Kravtsov, V., Alca´ıno, G., Marconi, G., & Alvarado, F. 2009, A&A, 497, 371Kravtsov, V., Alca´ıno, G., Marconi, G., & Alvarado, F. 2010, A&A, submittedMackey, A.D., Broby Nielsen, P., Ferguson, A.M.N., Richardson, J. C. 2008,ApJ681, 17LMarino, A. F., Villanova, S., Piotto, G. et al. 2008, A&A, 490, 625 lease give a shorter version with: \authorrunning and / or \titilerunning prior to \maketitle5