Beryllium and iron abundances of the solar twins 16 Cygni A and B

Red (signal-to-noise ratio of S/N similar to 1000 pixel(-1)) and ultraviolet (S/N greater than or similar to 100 pixel(-1)) Keck High Resolution Echelle Spectrograph (HIRES) spectra (R similar to 45,000 = 3 pixels) are used to derive the iron (Fe) and beryllium (Be) abundances in each of the solar twins 16 Cygni A and B. Self-consistent spectroscopic solutions yield, for 16 Cyg A and B, respectively, T-eff = 5795 +/- 20 and 5760 +/- 20 K, log g = 4.30 +/- 0.06 and 4.40 +/- 0.06, xi = 1.25 +/- 0.05 and 1.12 +/- 0.05 km s(-1), and [Fe/H] = 0.04 +/- 0.02 and 0.06 +/- 0.02. If Fe is used as a surrogate for metallicity, this represents an average metallicity of 11% +/- 5% above solar. These are in excellent agreement with other recent studies of this (wide) binary. Whereas it can be argued that no single study is conclusive, the consistent findings of these various studies offer compelling evidence that these stars have just barely supersolar metallicity, that 16 Cyg A is just hotter than the Sun, and that 16 Cyg B is just cooler. We have previously reported (based on Keck HIRES data) a difference in the lithium (Li) abundances of these stars of at least a factor of 4.5; for 16 Cyg A we detected a Li abundance of a factor of similar to 2 above solar, and for 16 Cyg B we placed a conservative upper limit of a factor of similar to 3 below solar. We detect Be in both stars and find that, if there is any difference between them, it must be much smaller-conservatively no more than 0.2 dex. Evidence suggests that solar-type stars deplete their surface Li abundance during the main sequence, a feat that the standard stellar evolution theory has, thus far, been unable to accomplish. Whatever physical mechanism depletes the surface Li abundance must create far less of a spread in the Be abundances than it does in the Li abundances. We find that our Li and Be results are consistent with the predictions of Yale models that include rotationally induced mixing driven by angular momentum loss. Our results provide no evidence for a small (similar to 0.05 dex) enhancement in the Be-9 abundance of the A component relative to the B component expected if the stars' Li abundance difference was due to accretion of planetary material by the A component. Given the errors, however, neither are we able to firmly preclude such a signature.