Constraints from stellar models on mixing as a viable explanation of abundance anomalies in globular clusters

Abstract

A detailed examination is made of the surface abundance changes (particularly of the CNO nuclei) which are predicted and observed to occur in evolving globular cluster (GC) stars. Discussed first are the standard stellar evolutionary sequences which have been constructed for a mass of 0.8 an d a metallicity of Z = 1CT 4 ([Fe/H] = -2.25), for comparison with observations of extremely metal-deficient cluster stars, and for a mass of 0.9 3K 0 and a metal content of Z = 6 X 10 3 ([Fe/H] = -0.49), for comparison with data from relatively metal-rich systems. These calculations illustrate the well-known problem that canonical models are completely inadequate insofar as being able to explain either the observed steep decline of the carbon abundance as a function of luminosity along the giant branches of M15 and M 92 or the CN-bimodality which is seen among giant stars of similar magnitude and color in several intermediate-metallicity and metal-rich GCs (e.g., NCG 362, 47 Tue).Then, on the premise that enhanced mixing of the envelope into the stellar interior is responsible for the observed abundances, partially-mixed models have been constructed in which artificially deep mixing is initiated at a wide variety of evolutionary phases and maintained during the subsequent evolution. The purpose of these calculations was not only to attempt to reproduce theoretically the abundance data but also to try to determine when the mixing could occur without producing accompanying side effects that violated other observational constraints. The present work shows, for instance, that the onset of deep mixing, of a sufficient degree to alter the surface composition in the requisite way, at any time between the zero-age main sequence (ZAMS) and the turnoff will induce a significantly larger color spread on the C-M plane than is generally permitted by recent CCD photometry. Indeed, in the case of 47 Tue, it is argued that the tightness of its C-M diagram, the reported good agreement between its observed luminosity function with that derived from canonical models, and the inferred small spread in the star-to-star helium abundances among the cluster horizontal-branch stars all provide persuasive evidence that unusual mixing in postmain-sequence stars does not occur in this system. Worthy of note, however, is the suggestion from our analysis that a mixing explanation for the CN-bimodality in GCs may be least problematical if the abundance anomalies arise during the very early evolutionary history of a star. What would seem to be required is that stars which presently show strong CN must have remained fully mixed as they initially contracted toward the ZAMS and somewhat beyond, until n( 12 C) ~ n( 14 N), and then evolved normally. However, even if stars can remain fully mixed for the time required, this possibility is not without its difficulties. Consequently, it is highly probable that some explanation other than mixing is responsible for the abundance anomalies which are seen in the bimodal-CN clusters. In contrast, the observed tr...

The lack of close binaries among hot horizontal branch stars in NGC 6752

“…As suggested by many in the literature (e.g. Vandenberg & Smith 1988;Sweigart 1997;D'Antona et al 2002), these abundance anomalies (either primordial, or due to mixing effects) can affect the RGB evolution and, consequently, have an impact on the properties of HB stars including temperatures, luminosities, gravities, and pulsation characteristics.…”

A Comparison between Carbon and Nitrogen Abundances of Bright Giants in the Globular Clusters M13, M10, and NGC 7006

“…This phenomenon is usually considered to be the result of progressive mixing to the stellar surface of CN(O)-processed material that has been exposed to protoncapture reactions in a region above the H-burning shell (e.g., Sweigart & Mengel 1979;Smith & Tout 1992;Charbonnel 1994Charbonnel , 1995Denissenkov & Weiss 1996). The standard first dredge-up event is not thought to be responsible, since neither is the convective envelope predicted by standard mixing-length theory deep enough nor the base temperature hot enough to produce the large observed drop in ½C=Fe abundance (e.g., VandenBerg & Smith 1988). Instead, some other process must be inferred that transports material between the base of the convective envelope and the close vicinity of the H-burning shell (e.g., Sweigart & Mengel 1979;Wasserburg et al 1995;Denissenkov & Tout 2000).…”

A search for late-type supergiants in the inner regions of the Milky Way

“…However, the fact that our stars most likely have a higher average metallicity than those of the solar neighbourhood and also those of the galactic center leads us to consider metallicity, rather than rotation, as the reason for the enhanced surface CNO element abundance. Higher metallicity implies higher opacity in the interior of the star, leading to a deeper penetration of the outer convective zone during the first dredgeup (Vandenberg & Smith 1988;Gratton et al 2000). In turn higher mass loss rates, enhanced by the higher metallicity, may also play a role by efficiently removing the outer envelope and rapidly make visible at the surface deep layers whose composition has been altered by CNO reactions (G. Meynet, priv.…”