Variable stars in one open cluster within the Kepler/K2-Campaign-5 field: M 67 (NGC 2682)††thanks: Based on observations collected with the Schmidt 67/92 Telescope at the Osservatorio Astronomico di Asiago, which is part of the Osservatorio Astronomico di Padova, Istituto Nazionale di Astrofisica. ††thanks: Light curves of variable stars available at http://groups.dfa.unipd.it/ESPG/aphn.html .
In this paper we continue the release of high-level data products from the multiyear photometric survey collected at the 67/92 cm Schmidt Telescope in Asiago. The primary goal of the survey is to discover and to characterise variable objects and exoplanetary transits in four fields containing five nearby open clusters spanning a broad range of ages.
This second paper releases a photometric catalogue, in five photometric bands, of the Solar-age, Solar-metallicity open cluster M 67 (NGC 2682). Proper motions are derived comparing the positions observed in 2013 at the Asiago’s Schmidt Telescope with those extracted from WFI@2.2m MPG/ESO images in 2000. We also analyse the variable sources within M 67. We detected 68 variables, 43 of which are new detection. Variable periods and proper-motion memberships of a large majority of sources in our catalogue are improved with respect to previous releases. The entire catalogue will be available in electronic format.
Besides the general interest on an improved catalogue, this work will be particularly useful because of: (1) the imminent release of Kepler/K2 Campaign-5 data of this clusters, for which our catalogue will provide an excellent, high spatial resolution input list, and (2) characterisation of the M 67 stars which are targets of intense HARPS and HARPS-N radial-velocity surveys for planet search.
keywords:techniques: photometric – stars: variables: general – binaries: general – open clusters and associations: individual: M 67 – proper motions
In 2015MNRAS.447.3536N we presented our multi-year, multi-wavelength photometric survey programme: “The Asiago Pathfinder for HARPS-N” (hereafter, APHN; PI: Bedin) aimed at characterising variable stars and transiting-exoplanet candidates in five open clusters (OCs). Originally, APHN was intended as a preparatory survey for the on-going searches of planets in OCs with the High Accuracy Radial velocity Planet Searcher for the Northern hemisphere (HARPS-N) mounted at the Telescopio Nazionale Galileo (TNG). The APHN survey has also recently acquired further interest, as four out of the five monitored OCs were chosen as targets for the Kepler extended mission K2111http://keplerscience.arc.nasa.gov/K2/ (2014PASP..126..398H).
In Paper I we analysed the two OCs M 35 and NGC 2158, for which we released atlases and stacked images. In this second paper we present the third of our target OCs: M 67 (NGC 2682). The OC M 67 has been the subject of many investigations. Among them: the search and study of variable stars (1991AJ....101..541G; 2002A&A...382..899S; 2002A&A...382..888V; 2003AJ....126.2954S; 2003AJ....125.2173S; 2006MNRAS.373.1141S; 2007MNRAS.377..584S; 2007MNRAS.378.1371B; 2008MNRAS.391..343P; 2009A&A...503..165Y), the determinations of proper motions and memberships (1977A&AS...27...89S; 1989AJ.....98..227G; 1993A&AS..100..243Z; 2008A&A...484..609Y; 2015AJ....150...97G), radial velocities (1986AJ.....92.1100M; 1986AJ.....92.1364M; 1990AJ....100.1859M; 1992PASP..104.1268M; 1994AJ....108.1828M; 2011A&A...526A.127P; 2015AJ....150...97G), the estimates of cluster parameters (age 4 Gyr, distance 1 kpc, metallicity [Fe/H]0, ; 1984AJ.....89..487J; 1986AJ.....92.1364M; 1987AJ.....93..634N; 1992AJ....103..151D; 1993AJ....106..181M; 1994A&AS..103..375C; 1996AJ....112..628F; 2003MNRAS.345.1015G; 2004PASP..116..997V; 2004MNRAS.347..101S; 2006A&A...450..557R; 2007A&A...470..585B; 2007AJ....133..370T; 2008A&A...489..677P; 2009ApJ...698.1872S; 2010A&A...511A..56P; 2011AJ....142...59J), the search for exoplanets candidates (2012A&A...545A.139P; 2014A&A...561L...9B), and the dynamical studies (2010A&A...513A..51B; 2012AJ....143...73P).
Our investigation is focused on finding new variable stars and extracting a complete astro-photometric catalogue (with improved proper motions and membership probabilities) in a region of degree, centred M 67.
The paper is organised as follows. In Section 2 we describe our data set, the data reduction, the extraction and the detrending of the light curves. In Section 3 we show the tools used for finding variable stars. Section 4 is dedicated to variable star detection, proper motions and membership probabilities computation, and colour-magnitude diagrams (CMDs) presentation. In Section 5 we describe the released electronic material. Finally, in Section 6 there is a summary of our work.
2 Observations, data reduction, light curves extraction and detrending
All images of the OC M 67 [()=()] were collected with the Asiago 67/92 cm Schmidt Telescope located on Mount Ekar (longitude E, latitude N, altitude 1370 m), that belongs to the Astronomical Observatory of Padova (OAPD), which is part of the Istituto Nazionale di Astrofisica (INAF). At the focus of the Schmidt telescope there is a SBIG STL-11000M camera, equipped with a Kodak KAI-11000M detector ( pixel, field of view: arcmin, pixel scale: 862.5 mas pixel). The characteristics of the telescope and the CCD are described in details in Paper I.
The OC M 67 is one of the four fields observed under the long-term observing programme APHN (PI: Bedin), aimed to characterise variable stars and transiting-exoplanet candidates in five OCs.
As in the case of M35 and NGC 2158 (Paper I), in the first observing season (2012) M 67 data were collected in white light (hereafter indicated with filter , where stands for ’None’), with exposure time of 120 s, and 60 s (during the almost-full moon nights); during the second (2013) and the third season (2014) we collected 180 s+15 s -filter and 180 s -filter images. Finally, during the fourth season (2015), the observations were carried out in -band (240 s+15 s) and -band (240 s). In table 1 we give a log of the observations, while in Fig. 1 we show the histograms of the number of images taken during all the observing seasons.
All images are stored in the INAF national archive in Trieste222http://ia2.oats.inaf.it/archives/asiago.
|Filter||# Images||Exp. Time||FWHM||Median FWHM|
For the data reduction, we used the software described in detail in Paper I. For each image we made a grid of spatially-varying, empirical point spread functions (PSFs), one for each of the 45 regions in which we divided the field of view (FOV); the software is similar to that developed by 2006A&A...454.1029A for the Wide Field Imager (WFI) mounted at the ESO / Max-Planck-Gesellschaft (MPG) 2.2m telescope.
For any location on the detector, the best PSF model is obtained by a bilinear interpolation of the four closest PSFs and it is used to measure the position and the flux of the stars in each image.
We transformed stellar positions in all the images into the reference frame of the best image333The “best image” is characterised by the minimum of the product between airmass and seeing. in filter (ID SC02906). For each filter, we obtained the photometric zero-points between the single images and the best image in that filter (ID: SC02906 for -filter, SC37438 for -filter, SC37437 for -filter, SC46181 for -filter, SC46722 for -filter).
We created a stacked, high-signal-to-noise ratio (SNR) image (‘stack’) of the field for each filter. Using the -filter stack, characterised by the highest SNR (Fig. 2), and the software used for single images, we derived an improved star list. We purged this star list from false detection using the qfit parameter (a diagnostic related to the quality of the PSF fit, 2008AJ....135.2114A) and the procedure described in 2014A&A...563A..80L. The purged star list contains 6905 sources, and we used it as the master star list (MSL) for the extraction of the light curves (LCs). The MSL contains stars with ; faint enough to reach the magnitude of the faintest white dwarfs (WDs) detected along the cooling sequence of M 67 by 2010A&A...513A..50B. The three bottom panels of Fig. 2 are centred on the three WDs shown in Fig. 2 of 2010A&A...513A..50B. The bottom-right panel of Fig. 2 shows the faintest WD () identified by the same authors. We extracted the -photometry of the 6905 sources in our catalogue by using the -stacked images. We calibrated the magnitudes by matching our catalogue to the Stetson Standard star catalogue (2000PASP..112..925S). We derived calibration equations by means of least squares fitting of straight lines using magnitudes and colours.
For the LCs extraction, we used the software developed and described in details in Paper I. Briefly, by using (i) the MSL catalogue, (ii) the PSFs, and (iii) the six-parameters linear transformations between the MSL and the single-image catalogues, for each target star within the MSL, the software extracts LCs in two parallel versions: a first one from the original images, and a second one from images where the neighbours close to the target star were PSF-fitted and subtracted. In both versions, our tool extracts the flux of the target star using PSF fitting and aperture photometry. As in Paper I, for aperture photometry we adopted a dynamical aperture that depends on the Full Width Half Maximum (FWHM), with radius =. Therefore, for each star in the MSL, four light curves are generated: PSF with/without neighbour-subtracted and aperture with/without neighbour-subtracted.
In order to remove residual systematic errors from the LCs, we followed the procedure used in Paper I and described in detail in 2014MNRAS.442.2381N. For each target star in each image, our algorithm computes local, weighted photometric zero-points using selected reference stars (generally the stars with the best rms at a given magnitude). The weights are a function of the relative on-sky position and of the magnitude difference between the target and the reference star. The software also computes a global zero-point correction, which —on average— provides worse residuals than the local zero-point correction, as also found in Paper I.
In Fig. 3 we show the photometric rms of the LCs derived from all the photometric methods (aperture with/without neighbours and PSF-fitting with/without neighbours). Even if the field of M 67 is relatively loose (if compared to the field analysed in Paper I) the photometry on images after neighbours subtraction is better than the photometry on original images, as the rms-scatter is lower and the rms-improvement is – on average – % in the case of aperture photometry, and % in the case of PSF-fitting photometry. Images were slightly de-focused to avoid saturation of main sequence (MS) turn off stars, therefore it was difficult to perfectly model the PSFs. This is also the reason why the rms PSF-fitting photometry is overall worse than the one associated to aperture photometry, as shown in Fig. 4.
3 Variable finding
In order to detect candidate variables in our field, we used three different algorithms: the Generalized Lomb-Scargle (GLS) periodogram (2009A&A...496..577Z), suitable for sinusoidal signals; the Analysis of Variance (AoV) periodogram (1989MNRAS.241..153S), useful to detect all variable types; the Box-fitting Least-Squares (BLS) periodogram (2002A&A...391..369K), effective for searching box-like dips in an otherwise flat or nearly flat LC, such as eclipsing binaries (EBs) and planetary transits. All the algorithms are part of the publicly available code VARTOOLS v.1.32444http://www.astro.princeton.edu/jhartman/vartools.html (2008ApJ...675.1254H). We used the output parameters associated to each algorithm for excluding the sources in our catalogue that have low probabilities to be variable in our data. In order to isolate the candidate variable stars, we used the procedure described in Paper I and summarised in Fig. 5.
From the histograms of the detected periods for all the LCs, we removed the spikes (saving the stars with high SNR, left panel of Fig. 5). The spikes are associated to spurious periods due to systematic effects, such as instrumental and atmospheric artefacts. In a second step, we selected by hand the candidate variables in the SNR versus period diagram (right panel of Fig. 5), and we visually inspected each of them.
We performed this procedure on neighbour-subtracted LCs, both for aperture and PSF photometries and for and LCs, identifying 68 real variable sources.
As in Paper I, we refined the periods using the following procedure: for each variable star, we normalised the and LC to zero, subtracting the 5-clipped median magnitude. Then, we merged the and normalised LCs, obtaining a LC with a temporal baseline of 764 days. We used again the VARTOOLS algorithms LS, AoV, and BLS to improve the period of the variable star. This procedure is useful only to improve the periods; it is not possible to extract any other scientific information from this normalised LC.
4 Variable stars and colour-magnitude diagrams
In our catalogue there are 6905 sources that cover a field-of-view of degree centred on M 67. In this catalogue we find 68 variables stars. All the variable stars are listed in Table 2; for each variable we provide the identification number (ID), the position, the period (if it is not irregular), and, when available, the magnitudes in bands, the membership probabilities as obtained in Sect. 4.1, and the radial velocities from 2015AJ....150...97G.
Of these 68 variable stars, 25 variable stars have already been classified in other photometric works (1991AJ....101..541G; 2002A&A...382..899S; 2002A&A...382..888V; 2003AJ....126.2954S; 2003AJ....125.2173S; 2006MNRAS.373.1141S; 2007MNRAS.377..584S; 2007MNRAS.378.1371B; 2008MNRAS.391..343P; 2009A&A...503..165Y) and/or in the General Catalogue of Variable Stars (GCVS, 2013A&AT...28...49S). Other variable stars listed in literature catalogues, but not found in our variable catalogue, are bright objects extremely saturated in our data (even in short exposures) or just outside the Schmidt FOV.
4.1 Proper motions and membership probabilities
We used stellar proper motions to separate cluster members and field stars. The approach adopted to compute stellar positional displacement is the same as in many other works, e.g., 2003AJ....126..247B; 2006A&A...454.1029A; 2008A&A...484..609Y; 2010A&A...513A..50B; 2015A&A...573A..70N, and 2015MNRAS.450.1664L. Using six-parameter local transformations and a sample of likely cluster members (for example MS stars), we computed the displacement between the stellar positions in two different epochs, after been transformed into a common reference system. As first epoch, we used M 67 -filter observations collected with the WFI mounted at the ESO/MPG 2.2 m telescope, on February 16th, 2000 (=2000.1). As second epoch we used the best 100 Schmidt -images collected during the 2013 observational run (2013.1). The time baseline for the proper motion measurements is 13.0 yr.
Since we used likely cluster members to compute the coefficients of the six-parameter linear transformations, the stellar displacements are computed relative to the cluster mean motion. Therefore, by construction, the cluster distribution in the vector-point diagram (VPD) is centred around (0,0), while field stars, that have a different motion relative to that of the cluster, lie in a different region of the VPD (see Fig. 6).
The membership probabilities (MPs) of the stars in the M 67 FOV have already been calculated using both astrometry (e.g. 1977A&AS...27...89S; 1989AJ.....98..227G; 1993A&AS..100..243Z; 2008A&A...484..609Y) and radial velocities (e.g 2008A&A...484..609Y; 2011A&A...526A.127P; 2015AJ....150...97G). Our final catalogue supplements the other works; we extracted the MPs in an homogeneous way for stars with in a region of arcmin.
There are different methods to derive stellar MPs. We chose that described by 1998A&AS..133..387B. Briefly, we derived the frequency function for both cluster () and field stars (). We assumed that cluster distribution is centred at (,) (0.00,0.00) mas yr with an intrinsic dispersion555The intrinsic dispersion is dominated by the positional errors and do not reflect the true intrinsic cluster dispersion. of (,) (0.65,0.71) mas yr. For field stars, we have (,) (9.34,1.10) mas yr and a dispersion of (,) (4.19,6.19) mas yr. We ignored the spatial distribution of the two components and assumed that there is not a correlation between them (correlation coefficient is set to 0). We excluded from our calculation poorly-measured-proper-motion stars. The membership probability is then computed as:
Panel (a) of Fig. 6 shows the available MPs as a function of the magnitude . The figure shows that we measured reliable MPs for stars with .
4.2 Colour-magnitude diagrams and light curves
In panel (b) of Fig. 6 we show the versus CMD of M 67. In panels (a) and (b), we plot in black the likely cluster member stars, selected according to their proper motions (i.e. the stars inside the red circles in the VPD of panels c), in azure the stars rejected and those are likely field stars, and with green dots the variable stars found in this work. In panel (d), using the same colour-code, we plot the positions of the stars in our MSL field of view.
In panels (e) there are four examples of variable stars with %. The LCs are normalised to the median magnitude. We plot in black the LC in -filter and in red that in -filter.
Panel (e) shows an eclipsing binary (ID #1840) that has never been photometrically observed, but that is in the sample of 2015AJ....150...97G. From spectroscopic observations, they found that this is a double lined (SB2) eclipsing binary and also a X-ray source (X35 in the ROSAT catalogue, 1998A&A...339..431B). In our data, we probably observe only the primary eclipse, that has a box-like shape with a depth of 0.02 mag.
5 Electronic material
The catalogue of all the sources in our MSL is electronically available. The catalogue contains the following information: Cols (1) and (2) are the J2000.0 equatorial coordinates in decimal degrees; Cols (3) and (4) the positions in pixel on the -filter stack; Cols (5)-(12) are the instrumental , and the calibrated magnitudes (when the magnitude is not available, it is flagged with -99.999); Col. (13) is the identification number (ID) of the star; Cols. (14) and (15) give the relative proper motions in mas yr along () direction (when it is not available, it is flagged with -999.9999); Col. (16) gives the membership probabilities (when it is not available, it is flagged with ).
Along with the MSL, we release the catalogue of all variable stars found in this work. An example of the catalogue is Table 2: Col. (1) is the identification number (ID) of the variable star in our MSL; Cols (2) and (3) are the J2000.0 equatorial coordinates in decimal degree; Col. (4) contains the periods in day (when the star is irregular, the period is -99); Cols (5)-(12) are the instrumental , and the calibrated magnitudes (when the magnitude is not available, it is flagged with -99.999); Col. (13) gives the membership probabilities (when it is not available, it is flagged with ); Col. (14) are the radial velocities measured by 2015AJ....150...97G (when the radial velocity measurement is not available, it is denoted as -1000). For each variable star, we also release the LCs.
Finally, we make the astrometrized stacks in filters electronically available.
In this work we present the photometric results for the third target (M 67) of the photometric survey of OC stars conducted with the 67/92 cm Schmidt telescope at Mount Ekar, Asiago. We analysed a total of 3707 images in , , , and white light (no filter), collected over 3.1 years. We used the algorithms described in 2015MNRAS.447.3536N to obtain a complete list of stars (6905 sources with magnitude ) and to extract, detrend, and analyse the corresponding LCs. We identify 68 variable stars (43 of which are new). Combining Schmidt best data with WFI@2.2m MPG/ESO images (collected on 2000), we derived the relative proper motions and the membership probabilities for a great number of stars in our MSL. We release two electronic catalogues: the catalogue of variable stars, (containing coordinates, periods , , , , , 2MASS magnitudes, membership probabilities, radial velocities) and the catalogue of all the sources in the Schmidt FOV (containing positions, , , , , , 2MASS magnitudes, proper motions, membership probabilities). The electronic material includes the , , , , and white light stacked images and light curves of the identified variable stars.
The OC M 67 is in the field of the K2 Mission-Campaign-5. The released catalogue of M 67 sources will be an excellent input-list for the extraction of LCs from K2 images. In this sense, our survey is preparatory to the analysis of K2 data, and complements (and extends in time) the light curves of the stars covered by K2.
We warmly thank the referee, Dr. R. Gilliland, for the prompt and
careful reading of our manuscript.
DN, ML, LRB, GP, AC, LB, and VG acknowledge PRIN-INAF 2012 partial funding under the project entitled: “The M 4 Core Project with Hubble Space Telescope”. DN and GP also acknowledge partial support by the Università degli Studi di Padova Progetto di Ateneo CPDA141214 “Towards understanding complex star formation in Galactic globular clusters”.