.451H The Astrophysical Journal Supplement Series, 30:451-490, 1976 April .30. © 1976. The American Astronomical Society. All rights reserved. Printed in U.S.A. 6ApJS. 197 EVOLVED STARS IN OPEN CLUSTERS Gretchen L. H. Harris* David Dunlap Observatory, Richmond Hill, Ontario Received 1974 September 16; revised 1975 June 18 ABSTRACT Radial-velocity observations and MK classifications have been used to study evolved stars in 25 open clusters. Published data on stars in 72 additional clusters are rediscussed and com- bined with the observations friade in this investigation to yield positions in the Hertzsprung- Russell diagram for 559 evolved stars in 97 clusters. Ages for the parent clusters were estimated from the main-sequence turnoff points, earliest spectral types, and bluest stars in the clusters themselves. The evolved stars were sorted into six age groups ranging from 4 x 106 yr to 4 x 108 yr, and the composite H-R diagram for each age group was then used to study the evolutionary tracks for stars of various masses. The observational results were found to be in reasonably good agreement with recent theoretical computations. The composite color-magnitude diagrams were found to be strikingly different from those of the rich open clusters in the Magellanic Clouds. At a given age the red giants in the Small Magellanic Cloud and the Large Magellanic Cloud clusters are brighter and bluer than their galactic counterparts. It is suggested that these effects may be accounted for by differences in metal abundance. Subject headings: clusters: open — galaxies: Magellanic Clouds — radial velocities — stars : evolution — stars : late-type — stars : spectral classification 1. INTRODUCTION Colour-Magnitude Diagrams (Hagen 1970) yielded 72 clusters younger than the Hyades in which member- Although open clusters are important tools in the ship judgment could be made by means of at least study of stellar evolution, their value is limited by the one of the above criteria. These data were supple- fact that most of them contain few evolved stars. This mented by new radial velocities and classification means that detailed comparisons between theoretical spectra for evolved stars in 25 open clusters. The and observational evolutionary tracks are possible for final list of 559 stars in 97 clusters was then used to only a small number of individual clusters. In ad- produce composite H-R diagrams for six age groups dition, the very sparseness of the evolved-star popula- 6 8 ranging from 5 x 10 yr to 5 x 10 yr. tion increases the potential influence of field stars in a given cluster color-magnitude diagram. II. THE OBSERVATIONS These problems can be alleviated by taking several clusters of a given age, each with good membership The stars to be observed in this program were information for the individual stars, and using them chosen from a sample of about 50 clusters for which to form a composite H-R diagram. With this in mind very little information was available. For practical an extensive survey of evolved stars in open clusters reasons the program was limited to stars brighter has been carried out. The ultimate aim of this pro- than Blim =+9.5 to +10.0. Whenever possible, gram was the production of a list of stars, confirmed preference was given to clusters with likely evolved as cluster members, for which ages could be deter- members which either appeared to be luminous from mined and composite H-R diagrams constructed. the cluster color-magnitude diagram or lay in unusual Since clusters younger than the Hyades are the ones or sparsely populated areas of that diagram. In most severely affected by problems related to small practice, the limit for inclusion in the observing numbers of stars, this investigation concentrated on program was often the star’s apparent magnitude. 8 clusters younger than ~5 x 10 yr. Further restric- Very few evolved-star candidates brighter than Mv = tions were imposed by the necessity of working with — 4 were found, and the final selection contained a clusters for which good membership data (spectral high percentage of G and K stars with — 3 < Mv < 0. classes, radial velocities, or proper motions) were Main-sequence stars were observed only to provide available in addition to reliable UBV photometry. better data on cluster radial velocities. An examination of the Atlas of Open Cluster The classification and radial-velocity spectrograms were obtained with the 91 cm telescope at the Cerro * Visiting Astronomer Cerro Tololo Inter-American Observatory, 1970 and 1971. The Cerro Tololo Inter-American Tololo Inter-American Observatory (CTIO) and the Observatory is operated by AURA, Inc., under contract with 61cm and 1.9 m telescopes of the David Dunlap the National Science Foundation. Observatory (DDO). Table 1 lists the telescopes used, © American Astronomical Society • Provided by the NASA Astrophysics Data System .451H 452 HARRIS .30. TABLE 1 Characteristic of the Spectrographs Used 6ApJS. Dispersion-1 Slit Width/Length Angular Slit Width 197 Telescope (Âmm ) (mm) (arcsec) Use Cn0 91cm 125 0.02/0.59 3 MK Classifications DDO 61 cm 112 0.025/0.6 2 MK Classifications CTIO 91 cm 62 0.006/0.3 1 Radial Velocities DDO 1.8 m 43 0.036/0.3 0.75 Radial Velocities the plate dispersions in Âmm-1, the projected slit The first is NGC 3114-59 (G8 III CN-), for which length and width in millimeters, the angular slit the Hagen classification is uncertain but not poor width in arcseconds, and comments as to whether enough to permit a temperature class earlier than G8. the plates were taken for purposes of classification However, the star appears to be CN-weak and should or radial velocity. All spectrograms were taken on be reexamined at a later date with the aid of CN Kodak Ila-O emulsion and developed in MWP2 strength standards. developer (Difley 1968). The second is NGC 6405-1 (BM Sco), a possible long-period variable (type SRd). The two temperature classifications given in Table 3 are essentially the same d) The Spectral Classifications but the luminosity class difference is definitely un- The southern classification spectrograms were ob- satisfactory. Keenan (19736) has indicated that this tained primarily during a 13 night run in 1970 May- star is somewhat inconsistent from plate to plate June with the Cassegrain spectrograph of the 61 cm and in some cases looks like a Ha instead of a lb. telescope at CTIO. A set of standard-star spectro- He feels the CN strength assignment is debatable, grams was obtained at normal, half-normal, and twice- but on the basis of several spectra would consider normal exposures. A comparison of the plates taken K2.5 Ib-II CN + 1 a satisfactory compromise at this on the two CTIO observing runs showed that the point. In contrast, the single plate of NGC 6405-1 focus for the later set of plates was somewhat sharper. obtained for this investigation shows no indication of The earlier run of plates is quite satisfactory, how- having a higher luminosity than class II or stronger ever, and the standards are mainly at that focus. Since than normal CN. this does present a slight problem in comparison with the 1971 spectra, classifications based on the b) The Radial Velocities later observing run have been indicated with an asterisk in Table 2 as described below. All of the southern radial-velocity spectrograms Although a few standard-star spectra were taken were obtained in the course of a 19 night run at CTIO on the northern observing run to act as a check on the in 1971 January-February. The northern radial spectrograph consistency, the actual classification velocity spectra were taken during the winter of 1971- of the northern stars was done with the aid of an 1972 with the Cassegrain spectrograph of the DDO almost complete standard-star file lent by Garrison. 1.9 m telescope. Effective wavelengths used for the All of the northern spectra were taken with a grating stellar and reference lines were those given by Batten spectrograph designed by Garrison for the DDO 61 et al. (1971) for a dispersion of 60 Âmm-1, supple- cm telescope in Richmond Hill. mented by wavelengths established by Rice (1966) The final classifications for the 73 stars are listed in for the early-type stars. Table 2. Stars will often be referred to in this paper Stars chosen for the radial-velocity program were by their cluster identifications, such as Cr 140-15. either evolved stars whose spectral classifications The references to the sources of these numbers are implied cluster membership or upper-main-sequence given in Appendix A. Only one spectrogram was cluster members. Radial velocities for probable main- taken for each program star unless the first was poor sequence cluster stars were needed to establish a and time was available for another. All spectrograms cluster velocity against which the evolved star measure- were classified during two periods several months ments could be compared. Since the percentage of apart. Uncertain or inconsistent classifications were spectroscopic binaries is likely to be high for upper- then reexamined before a final spectral class was main-sequence stars, at least three or four stars were assigned. chosen to define an approximate cluster velocity. A few stars in this investigation were studied simul- Whenever possible, three or four plates were taken of taneously by others. The comparisons of Hagen with each main-sequence star and each evolved star. In Frye, MacConnell, and Humphreys (1970; FMH) addition to the program stars, IAU standard velocity and Keenan (1972; K) are given in Table 3. In stars (Pearce 1955) were observed every night. general, the agreement is satisfactory and well within Table 4 gives the results for the standard velocity the expected accuracy of the MK system.