BERYLLIUM AND ALPHA-ELEMENT ABUNDANCES IN A LARGE SAMPLE OF METAL-POOR STARS

The light elements, Li, Be, and B, provide tracers for many aspects of astronomy including stellar structure, Galactic evolution, and cosmology. We have made observations of Be in 117 metal-poor stars ranging in metallicity from [Fe/H] = -0.5 to -3.5 with Keck I/HIRES. Our spectra are high resolution (similar to 42,000) and high signal to noise (the median is 106 per pixel). We have determined the stellar parameters spectroscopically from lines of Fe I, Fe II, Ti I, and Ti II. The abundances of Be and O were derived by spectrum synthesis techniques, while abundances of Fe, Ti, and Mg were found from many spectral line measurements. There is a linear relationship between [Fe/H] and A(Be) with a slope of +0.88 +/- 0.03 over three orders of magnitude in [Fe/H]. We find that Be is enhanced relative to Fe; [Be/Fe] is +0.40 near [Fe/H] similar to -3.3 and drops to 0.0 near [Fe/H] similar to -1.7. For the relationship between A(Be) and [O/H], we find a gradual change in slope from 0.69 +/- 0.13 for the Be-poor/O-poor stars to 1.13 +/- 0.10 for the Be-rich/O-rich stars. Inasmuch as the relationship between [Fe/H] and [O/H] seems robustly linear (slope = +0.75 +/- 0.03), we conclude that the slope change in Be versus O is due to the Be abundance. Much of the Be would have been formed in the vicinity of Type II supernova (SN II) in the early history of the Galaxy and by Galactic cosmic-ray (GCR) spallation in the later eras. Although Be is a by-product of CNO, we have used Ti and Mg abundances as alpha-element surrogates for O in part because O abundances are rather sensitive to both stellar temperature and surface gravity. We find that A(Be) tracks [Ti/H] very well with a slope of 1.00 +/- 0.04. It also tracks [Mg/H] very well with a slope of 0.88 +/- 0.03. We have kinematic information on 114 stars in our sample and they divide equally into dissipative and accretive stars. Almost the full range of [Fe/H] and [O/H] is covered in each group. There are distinct differences in the relationships of A(Be) and [Fe/H] and of A(Be) and [O/H] for the dissipative and the accretive stars. It is likely that the formation of Be in the accretive stars was primarily in the vicinity of SN II, while the Be in the dissipative stars was primarily formed by GCR spallation. We find that Be is not as good a cosmochronometer as Fe. We have found a spread in A(Be) that is valid at the 4 sigma level between [O/H] = -0.5 and -1.0, which corresponds to -0.9 and -1.6 in [Fe/H].