A&A 509, A17 (2010)

The first search for variable stars in the open cluster NGC 6253 and its surrounding field^,

F. De Marchi^1,2 - E. Poretti³ - M. Montalto^4,5 - S. Desidera⁶ - G. Piotto¹

1 - Dipartimento di Astronomia, Università di Padova, Vicolo dell'Osservatorio 2, 35122 Padova, Italy
2 - Dipartimento di Fisica, Università di Trento, Via Sommarive 14, 38123 Povo (TN), Italy
3 - INAF - Osservatorio Astronomico di Brera, Via E. Bianchi 46, 23807 Merate (LC), Italy
4 - Universitaets-Sternwarte der Ludwig-Maximilians-Universitaet, Scheinerstr. 1, 81679 Muenchen, Germany
5 - Max-Planck-Institute for Extraterrestrial Physics, Giessenbachstr., 85748 Garching bei Muenchen, Germany
6 - INAF - Osservatorio Astronomico di Padova, Vicolo dell'Osservatorio 5, 35122 Padova, Italy

Received 22 May 2009 / Accepted 30 September 2009

Abstract
Aims. This work presents the first high-precision variability survey in the field of the intermediate-age, metal-rich open cluster NGC 6253. Clusters of this type are benchmarks for stellar evolution models.
Methods. Continuous photometric monitoring of the cluster and its surrounding field was performed over a time span of ten nights using the Wide Field Imager mounted at the ESO-MPI 2.2 m telescope. High-quality timeseries, each composed of about 800 datapoints, were obtained for 250 000 stars using ISIS and DAOPHOT packages. Candidate members were selected by using the colour-magnitude diagrams and period-luminosity-colour relations. Membership probabilities based on the proper motions were also used. The membership of all the variables discovered within a radius of 8 from the centre is discussed by comparing the incidence of the classes in the cluster direction and in the surrounding field.
Results. We discovered 595 variables and we also characterized most of them providing their variability classes, periods, and amplitudes. The sample is complete for short periods: we classified 20 pulsating variables, 225 contact systems, 99 eclipsing systems (22  Lyr type, 59  Per type, 18 RS CVn type), and 77 rotational variables. The time-baseline hampered the precise characterization of 173 variables with periods longer than 4-5 days. Moreover, we found a cataclysmic system undergoing an outburst of about 2.5 mag. We propose a list of 35 variable stars as probable members of NGC 6253.

Key words: starspots - stars: statistics - stars: variables: general - binaries: eclipsing - novae, cataclysmic variables - open clusters and associations: individual: NGC 6253

1 Introduction

NGC 6253 and NGC 6253 are the only open clusters whose metallicities above [Fe/H] = +0.3 were confirmed by spectroscopic analyses (Carretta et al. 2000,2007; Sestito et al. 2007). Therefore, these clusters are of special interest in several fields, e.g., as benchmarks for stellar evolution and stellar population models and as targets for the search for extrasolar planets. We observed both clusters in the framework of our project looking for transiting planets in super-metal-rich open clusters. The results obtained on NGC 6791 were presented by Montalto et al. (2007).

We also performed a 10-night observing campaign on NGC 6253 with the same purposes as for NGC 6791. In the first paper based on our new investigation, Montalto et al. (2009) obtained broad band photometry and astrometry for 187 963 stars within 30 arcmin from the cluster. Images from ESO archive (Momany et al. 2001) were also used to derive relative proper motions and then distinguish between field stars and cluster members. The availability of the astrometric cluster memberships and the photometric quality of the new data allowed new, independent determinations of the cluster's main parameters. Indeed, the determinations of the NGC 6253 parameters are affected by larger uncertainties because of the cluster's projection toward a very rich stellar field fairly close to the galactic centre (l = 335.46 deg, b = -6.25 deg). Systematic differences in the photometric calibrations of different datasets have been found (Sagar et al. 2001; Anthony-Twarog et al. 2007; Twarog et al. 2003; Bragaglia et al. 1997; Piatti et al. 1998). In this paper we adopt the values of the distance modulus and of the reddenings obtained by Montalto et al. (2009) using the technique of the isochrone fitting, i.e., (m-M)_V = 11.68  0.10 mag, E(B-V) = 0.15  0.02 mag and E(V-I) = 0.25  0.02 mag. These values are also consistent with a weighted mean of all the determinations. The cluster age is about 3.5 Gyr (Montalto et al. 2009).

Our project gives the possibility of studying stellar variability in super-metal-rich stars using high-quality data (De Marchi et al. 2007). Since no variability survey on NGC 6253 has previously been performed, we characterize the variable stars in NGC 6253 and in its surrounding field for the first time. To do that, we started from the new findings and calibrations obtained by Montalto et al. (2009) so we refer the reader to that paper for a more detailed explanation of the methodologies applied to determine the properties and the fundamental parameters of the cluster.

Table 1:   The observation log for each night and limits of the field of view.

2 Observations and data reduction

NGC 6253 was observed for 10 consecutive nights (from June 13, 2004 to June 22, 2004) using the Wide-Field Imager (WFI) mounted at the ESO-MPI 2.2 m telescope, La Silla, Chile. The WFI instrument includes a mosaic of eight 2k  4k CCDs. The pixel scale is 0.238 arcsec/pixel. In total, 45.3 h of observation were collected, mainly in the R filter. A few deep images in the B, V, and I filters were also acquired to construct colour-magnitude diagrams (CMDs), along with a standard field to allow the calibration of the data. In total 918 images of the cluster were obtained, with a mean exposure time of 178 s. Table 1 reports the journal of observations and Fig. 1 shows a WFI image of NGC 6253. Since the the size of each chip is 8 in right ascension and 16 in declination, we centered the cluster on one chip to minimize the loss of stars between chips. Observations and data reduction to derive the calibrated photometry and the CMDs of the cluster are described in more detail in Montalto et al. (2009). The procedure to derive the light curves uses both ISIS 2.2 (Alard 2000; Alard & Lupton 1998) and DAOPHOT II (Stetson 1998) packages, as described in Montalto et al. (2007).

The length of the observing nights (more than 0.32 d in 7 cases and more than 0.40 d in 5 cases, see Table 1) reduced the height of the aliases situated at 1 d^-1 from the central peak down to below 60% of the power (Fig. 2, upper panel). Moreover, the light curves are very dense and their shape clearly defined on each night (Fig. 2, middle panel). Both these facts made the period detection quite straightforward, not only in the case of high-amplitude variables, but most of time also for small-amplitude, short-period variable stars.

As can be noted in Fig. 1, our survey covers a much larger field of view than the previous ones (6.3  6.3 arcmin² by Bragaglia et al. 1997, 3.8  3.8 arcmin² by Piatti et al. 1998). We could also identify new variable stars in a wide part of the surrounding field. The ISIS 2.2. and DAOPHOT II packages returned a photometric precision well below 0.01 mag in the magnitude range . A plot of the standard errors of the mean magnitudes in different filters is shown in Fig. 1 in Montalto et al. (2009). Stars brighter than the turn-off magnitude (V = 14.5) are saturated in our photometry and cannot be studied. In particular, this constraint hampers the study of the variability of the blue stragglers, as performed by De Marchi (2008) in the more favourable case of NGC 6791.

3 Cluster membership

NGC 6253 is a relatively small cluster, but Bragaglia et al. (1997) noticed the necessity of moving 8 from the cluster centre to find a legitimate external field. We followed this prescription and we adopted the centre coordinates given by Bragaglia et al. (1997).

The measured stars are indicated by small points, the 150 stars with membership probability (hereafter MP) greater than 90% are highlighted with larger black points. MPs were calculated in Montalto et al. (2009) following the approach proposed by Vasilevskis et al. (1958):

where and are the distribution of cluster and field stars in the diagram of the proper motions, respectively. These distributions are typically represented as Gaussian functions. The distribution of the cluster stars has a narrow peak centered at , while the distribution of field stars is much broader. For the given candidate member, the calculation of the MP was performed by selecting a surrounding sample of a 2.5 mag range centered on the candidate's position. In such a way the local sample stars compensate for the effect of a magnitude dependence of the cluster-to-field star ratio. When constructing a   diagram, the stars belonging to the cluster occupy a well-defined region (see Fig. 4 in Montalto et al. 2009). We require the probable member clusters to have at V=12.5 and at V=18.0.

Since the determination of the MPs is a differential process and the cluster is almost completely included in chip 2, the MPs are reliable only for stars belonging to this chip and brighter than V=18. Looking at the distribution of the stars with a high MP we can infer that some members of the cluster might also be present in chips 1 and 3.

4 The variable stars

4.1 Detection

The ISIS 2.2 and DAOPHOT II packages allowed us to extract the first list of suspected variable stars from the full database of 250 000 timeseries. This list was validated and shortened by calculating the parameters related to the reduction of the initial variance obtained by introducing trial periodic terms. These parameters are the reduction factor (Vanícek 1971) and the coefficient of spectral correlation (Ferraz-Mello 1981). As in the case of NGC 6971 (De Marchi et al. 2007), we could separate short- and long- period variable stars by introducing a parameter that is more sensitive to the night-to-night variations. Tests on the significance of the detected periodicities (e.g., signal-to-noise ratio above 4.0 in amplitude) allowed us to get a more defined sample of real variable stars. A few objects whose variability appears to stem from photometric artefacts (e.g. eclipse-like features occurring exactly at the same time on the second night) were removed from the list. These spurious photometric effects are usually corrected when applying to the light-curve algorithms such as the one developed by Tamuz et al. (2005). However, we noticed that the application of this algorithm degrades the precision of the variable star photometry. Therefore, being interested in much greater light variations than the tiny photometric effect of a planetary transit, we decided to analyse the light curves before applying the algorithm.

We identified 595 variable stars at the end of our process, whose timeseries are composed of about 800 datapoints. To identify them we used the five-digit number assigned by our customized package package, followed by the number of the chip that the star belongs to. The timeseries are available at the ``Centre de données astronomiques de Strasbourg'' (CDS).

Table 2:   Inventory of the variables found in NGC 6253 and its surrounding area.

4.2 Classification

The timeseries of the 595 variable stars were analysed in frequency by using the least-squares iterative sine-wave search (Vanícek 1971) and the Phase Dispersion Minimization (Stellingwerf 1978) methods. The periods were refined by means of a least-squares procedure (MTRAP, Carpino et al. 1987); their error bars are in the range 1-6  10^-5 d. The bottom panel of Fig. 2 shows the distribution of the standard deviations of the least-squares fits, indicating a median precision of 0.015 mag.

We could show amplitudes of light variability down to the 0.01 mag level. At this level, rotational variables could be separated from pulsating variables on the basis of the period values and of the Fourier parameters alone (Poretti 2001). On the other hand, it is very difficult to disentangle rotational from eclipsing variables. To distinguish rotational variables from contact binaries, we referred to the degree of asymmetry of the double-wave light curves and to the occurrence of the minima at phases 0.00 and 0.50. Of course, we cannot rule out that a small fraction of the variables classified as rotational variables might be actually contact systems showing grazing eclipses or viceversa.

We considered two classes of rotational variables, RO1 and RO2 stars. RO1 stars show a light curve characterized by a single wave, which is often asymmetrical. RO2 stars show a more complicated curve composed of two waves having unequal amplitude and duration. This light curve is comes from two (groups of) spots located at different latitudes that remain visible to the observer during different fractions of the rotational period. In some cases these spotted stars are observed in eclipsing systems, the so-called RS CVn variables. Other cases of eclipsing systems are contact (W UMa variables, EW), semi-detached ( Lyr variables, EB), and detached systems (Algol variables, EA) binaries. In some cases it was very difficult to distinguish between EW system showing grazing eclipses and rotational variables. We also identified three different classes of pulsating variables, i.e., RR Lyr,  Sct and high-amplitude  Sct (HADS) stars. In both cases, eclipsing binaries and pulsating variables, the very good spectral window (Fig. 2) made the period detection quite straightforward. On the other hand, defining the periods longer than 4-5 d was not easy. In particular it was impossible for periods longer than 10 d and we simply classify these stars as long period (LON) variables. These stars are mostly rotational variables. The summary classification of the entire sample is reported in Table 2. Tables A.1-A.8 list the members of each class giving the identifier in the Montalto et al. (2007) catalogue, the coordinates, the photometry, the epoch of maximum or minimum brightness (HJD-2 453 100), the period, the amplitude, the distance, and the MP values. Uncertain MP values (stars with V>18, often close to the chip borders) are marked with an asterisk. The catalogue of the light curves of the periodic variables is available at the CDS.

We paid particular attention to the variables located within 8from the cluster centre; in any cases no variable with MP larger than 50% was found at a greater distance. If the MP is not available (stars near the edges of chip 2 or stars fainter than V=18), the membership is estimated from their location on the B-V vs. V and V-I vs. V CMDs. Moreover, for pulsating variables and contact binaries with unknown membership, our conclusions are based on the applications of the usual period-luminosity (P-L) and period-luminosity-colour (P-L-C) relations.

4.3 Pulsating variables

The only pulsating variable located at less than 8 from the centre is the RRab star 10540_2. Its MP is quite high, but it is clearly too faint (V=17.39) to belong to the cluster. Amongst the other four RR Lyr stars, 15578_7 is a new galactic Blazhko variable.

Twelve variables show an amplitude smaller than 0.06 mag; since they have a very short period (less than 0.10 d), we can rule out their being rotational variables. All the B-V values except one range from 0.42 to 0.79 mag, mostly between 0.50 and 0.63. This interval, taking the reddening E(B-V) = 0.15 mag into account, suggests their classification as  Sct stars. The remaining low-amplitude variable shows (B-V)=0.084: it probably belongs to the  Cep class.

Four variables show a larger amplitude (more than 0.09 mag) and the asymmetric shape of the light curve typical for high-amplitude  Sct stars. By using the new period-luminosity relation derived by Poretti et al. (2008), no doubt is left on the fact that all these variables do not belong to NCG 6253. Finally, none of the pulsating variables is a member of NGC 6253, since they are all located well beyond the cluster.

4.4 Contact binaries

We have 50 contact binaries located within the 8 radius and the (B-V) and (V-I) colours are both available for 44 binaries, while only the (V-I) colours are available for 6 of them. For these stars it is possible to apply the P-L-C relations given by Rucinski (2003) and compare the resulting distance moduli with that of the cluster, obtained by isochrone fitting (Montalto et al. 2009). The errors on the distance moduli are calculated by considering the uncertainties on the mean B-V and V-I colours. It must also be taken into account that, since our light curves are in the R band, we cannot know the exact value  of the magnitude at maximum brightness required by the Rucinski (2003) calibrations. We estimated  as ; i.e., we assumed that the colour of these binary systems does not change during the orbital period since the components have a very similar temperature.

An estimate of the membership of the objects can be obtained using the CMDs (Fig. 3, top row), the MPs based on the proper motions, and the two P-L-C relations (Fig. 3, bottom row). The fiducial lines shown in the CMDs are obtained by selecting 10-15 points at different magnitudes along the Main Sequences, both from the observations described here and from Bragaglia et al. (1997). We could select a list of 13 candidate members for which one of the above criteria is satisfied (Table A.9).

We note that 23188_2 and 10853_2 satisfy all the membership criteria and then are very likely cluster members. The position in the CMDs and the P-L-C relations also suggest that 01015_2 is a cluster member, but this hint is not supported by the MP, which is very small. The case of 09268_2 is the opposite: it has also a fairly large MP, but the other indicators suggest that is is more probably located between the Sun and the cluster. Unfortunately, none of the remaining cases gives us enough confidence on a cluster membership.

We can tackle the problem of cluster membership in an indirect way. In the surrounding field we found 175 EW binaries, with an incidence of 0.21 EW arcmin^-2. Therefore, we should have 42 field EW-stars superposed on the cluster. Since we found 50 stars (Table 2), the excess is only marginally significant. We have only two well-established memberships; therefore, we can reasonably estimate that very few contact binaries (up to six) among the remaining 11 candidates listed in Table A.9 actually belong to NGC 6253. This clue is confirmed by the candidates do not match the photometric criteria very well (Table A.9). In NGC 6791 we found three well-established and five likely EW-members (De Marchi et al. 2007), i.e., similar countings. The surveys of the two clusters are complete both at the magnitude and at the periods of the EW binaries. The two clusters have a different stellar content, since NGC 6253 has about 500-1000 members (Montalto et al. 2009), and NGC 6791 about 4900  1000 (De Marchi 2008). The similarity between EW countings in the two clusters supports the hypothesis of an anticorrelation between the frequency of binaries and the richness of the host cluster (Kaluzny & Rucinski 1995).

Among the non-member contact binaries, we note that 00441_4 has a period of 0.21002 d, shorter than the shortest contact binary found in the ASAS database (P = 0.217811 d, NGC 6253, Rucinski 2007) and very similar to the binary with the shortest period known (P = 0.2009 d, NGC 6253, Weldrake & Bayliss 2008).

4.5 Semi-detached and detached systems

The sample of the semi-detached and detached systems within 8 is composed of five EA, two EB, and two RS CVn stars (lower part of Table A.9). Their periods are shorter than 2.3 d. The star 26902_2 has a high MP, and it is the only case for which we can be very confident about its membership, also confirmed by the positions in the CMDs (Fig. 4, left panels). On the basis of the same criteria, 10340_2 is another probable member. On the other hand, the MP value rules out the membership of 00145_2. No firm conclusion on the membership can be drawn on the other cases.

4.6 Rotational and long-period variables

A great number of the new variables discovered in our survey shows the single (RO1) or double (RO2) wave light curves typical of rotational effect. The 10-d time baseline allowed us to detect all the variables with rotational periods shorter than 4-5 days. Other variables show an evident night-to-night variability, but we cannot infer any reliable value for the period. These variables are probably long-period ones (LON).

By adopting the same criteria as used in other cases, we selected the RO1 (14 stars), RO2 (2 stars), and LON (16 stars) candidate members of NGC 6253 (Table A.10). Figure 4 plots the CMDs with the positions of all the rotational variables within 8 from the centre and the positions of the long-period variables highlighted (middle and right panels, respectively).

We discovered 27 variables in the 8 radius from the centre and 16 of them can be considered candidate members on the basis of the positions on the CMDs and of the MPs (upper part of Table A.10). The stars 10042_2, 11077_2, and 06387_2 have a large MP and also a position on the CMDs compatible with cluster membership. We count 57 rotational variables in the surrounding field, i.e., an occurrence of 0.06 star arcmin^-2. This would imply an estimate of 12 field rotational variables along the line of sight of NGC 6253. There is a significant difference between the expected and the observed number of rotational variables, and we can infer that several selected candidate members (up to 15) actually belong to the cluster. Considering the short periods of these stars and the old age of the cluster, it is likely that the rotational variables that are cluster members are close, tidally locked binaries.

In the same manner, we can estimate 32 LON field variables superposed to NGC 6253. In turn this means that up to 9 out of the 41 LON variables discovered in the 8 radius can be considered members of the cluster. These 9 stars should be found among the 16 candidate members listed in Table A.10.

4.7 A new cataclysmic variable

The U Geminorum variable 15877_2 is located at 6.8 from the cluster centre, but unfortunately its MP is not available (Table A.10). In the light curve, the scatter at the quiescence phase suggests some photospheric activity, but no periodicity is detected by analysing these measurements. Such a phase lastes the first 8 days of our survey. After that, its brightness in the R-band increases by about 2.5 mag (Fig. 5). The maximum is not observed because it occurred in daytime. The star 15877_2 appears to be similar to the U Geminorum variable 06289_9, classified as a member of NGC 6791 (De Marchi et al. 2007).

5 Conclusions

In this paper we have described the first search for variable stars in the open cluster NGC 6253. Since the membership probabilities based on the proper motions are not reliable for stars with V>18, only a few variables could be confirmed directly as cluster members. However, the comparison with the number of contact binaries and rotational variables (both short and long periods) found in a large area surrounding the cluster allowed us to estimate the incidence of these variables within the cluster, too. On the basis of these considerations we propose 35 members of NGC 6253 within the sample of variable stars, though new observations are needed to identify some of them in an unambiguous way.

The class of main-sequence rotational variables is the most numerous, as observed in the surrounding field. On the basis of similar observing campaigns, we found the same number of contact binaries in NGC 6253 as were previously found in NGC 6791, thus confirming the anticorrelation between the frequency of binaries and the richness of the cluster (Kaluzny & Rucinski 1995). This anticorrelation is similar to the one found between the frequency of blue stragglers and the total magnitude of the host cluster. Both these facts can lead back to the important effects caused by mass loss in the evolution and in the history of the dynamics of open clusters (De Marchi et al. 2006; Davies et al. 2004).

We discovered a new eruptive variable in NGC 6253. A single outburst was observed, so we cannot infer any physical characteristic of the system. Since we made the same discovery in NGC 6971 (De Marchi et al. 2007), it seems that continuous surveys on a few nights are very effective in finding these rare and interesting objects.

Acknowledgements

This work was funded by COFIN 2004 ``From stars to planets: accretion, disk evolution and planet formation'' by MIUR and by PRIN 2006 ``From disk to planetary systems: understanding the origin and demographics of solar and extrasolar planetary systems'' by INAF. We thank the anonymous referee for careful reading and useful suggestions, and J. Vialle for checking the English form.

References

Appendix A: Tables

Table A.1:   Pulsating variables.

This Appendix includes the tables listing all the variables discovered in our survey of NGC 6253 and its surrounding field. The epochs of maximum or minimum brightness are expressed as HJD-2 453 100 in the columns  and  .

1.

Pulsating variables: Table A.1;

2.

EW-type variables: Table A.2;

3.

EA-type variables: Table A.3;

4.

EB-type variables: Table A.4;

5.

RS-CVn variables: Table A.5;

6.

rotational single-wave variables: Table A.6;

7.

rotational double-wave variables: Table A.7;

8.

long-period variables: Table A.8.

The binary systems considered as candidate members of NGC 6253 are listed in Table A.9. The rotational and long-period variables considered as candidate members of NGC 6253 are listed in Table A.10.

Table A.2:   Contact binaries, W Ursae Maioris (EW) systems.

Table A.3:   Detached systems, Per (Algol) type (EA) systems.

Table A.4:   Semi-detached systems, Lyr type (EB) systems.

Table A.5:   RS CVn variables, binary systems showing stellar activity in one or both components.

Table A.6:   Single-wave rotational variables (RO1).

Table A.7:   Double-wave rotational variables (RO2).

Table A.8:   Long-period variables.

Table A.9:   List of binary systems (contact, detached and semi-detached binaries) located at less than 8 from the centre which are considered as candidate members of NGC 6253.

Footnotes

... field

Based on observation made at the European Southern Observatory, La Silla, Chile, Proposal 073.C-0227.

...

Timeseries and light curves are available in electronic form at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/509/A17

Copyright ESO 2010

All Tables

Table 1:   The observation log for each night and limits of the field of view.

Table 2:   Inventory of the variables found in NGC 6253 and its surrounding area.

Table A.1:   Pulsating variables.

Table A.2:   Contact binaries, W Ursae Maioris (EW) systems.

Table A.3:   Detached systems, Per (Algol) type (EA) systems.

Table A.4:   Semi-detached systems, Lyr type (EB) systems.

Table A.5:   RS CVn variables, binary systems showing stellar activity in one or both components.

Table A.6:   Single-wave rotational variables (RO1).

Table A.7:   Double-wave rotational variables (RO2).

Table A.8:   Long-period variables.

Table A.9:   List of binary systems (contact, detached and semi-detached binaries) located at less than 8 from the centre which are considered as candidate members of NGC 6253.

All Figures

---

Copyright ESO 2010