The chemical composition contrast between M3 and M13 revisited: New abundances for 28 giant stars in M3

We report new chemical abundances of 23 bright red giant members of the globular cluster M3, based on high-resolution (R similar to 45,000) spectra obtained with the Keck I telescope. The observations, which involve the use of multislits in the HIRES Keck I spectrograph, are described in detail. Combining these data with a previously reported small sample of M3 giants obtained with the Lick 3 m telescope, we compare metallicities and [X/Fe] ratios for 28 M3 giants with a 35-star sample in the similar-metallicity cluster M13, and with Galactic halo field stars having [Fe/H]< -1. For elements having atomic number A >= A(Si), we derive little difference in [X/Fe] ratios in the M3, M13, or halo field samples. All three groups exhibit C depletion with advancing evolutionary state beginning at the level of the red giant branch "bump,'' but the overall depletion of about 0.7 - 0.9 dex seen in the clusters is larger than that associated with the field stars. The behaviors of O, Na, Mg, and Al are distinctively different among the three stellar samples. Field halo giants and subdwarfs have a positive correlation of Na with Mg, as predicted from explosive or hydrostatic carbon burning in Type II supernova sites. Both M3 and M13 show evidence of high-temperature proton-capture synthesis from the ON, NeNa, and MgAl cycles, while there is no evidence for such synthesis among halo field stars. But the degree of such extreme proton-capture synthesis in M3 is smaller than it is in M13: the M3 giants exhibit only modest deficiencies of O and corresponding enhancements of Na, less extreme overabundances of Al, fewer stars with low Mg and correspondingly high Na, and no indication that O depletions are a function of advancing evolutionary state, as has been claimed for M13. We have also considered NGC 6752, for which Mg isotopic abundances have been reported by Yong et al. Giants in NGC 6752 and M13 satisfy the same anticorrelation of O abundances with the ratio (Mg-25 + Mg-26)/Mg-24, which measures the relative contribution of rare to abundant isotopes of Mg. This points to a scenario in which these abundance ratios arose in the ejected material of 3 - 6 M-. cluster stars, material that was then used to form the atmospheres of the presently evolving low-mass cluster stars. It also suggests that the low oxygen abundance seen among the most evolved M13 giants arose in hot bottom O-to-N processing in these same intermediate-mass cluster stars. Thus, mixing is required by the dependence of some abundance ratios on luminosity, but an earlier nucleosynthesis process in a hotter environment than giants or main-sequence stars is required by the variations previously seen in stars near the main sequence. The nature and the site of the earlier process is constrained but not pinpointed by the observed Mg isotopic ratio.