Can surface oxygen abundances of red giants be explained by the canonical mixing theory?

Extensive oxygen abundance determinations were carried out for 239 late-G/early-K giant stars of 1.5-5 M_sun by applying the spectrum-fitting technique to O I 7771-5 and [O I] 6300/6363 lines based on the high-dispersion spectra in the red region newly obtained at Okayama Astrophysical Observatory. Our main purpose was to clarify whether any significantly large (<~0.4-0.5 dex) O-deficit really exists in these evolved stars, which was once suspected by Takeda et al. (2008, PASJ, 60, 781) from the analysis of the [O I] 5577 line, since it (if real) is inexlainable by the current theory and may require the necessity of special non-canonical deep mixing in the envelope. We found, however, that the previous [O/H]_5577 results (differential abundances relative to the sun) were systematically underestimated compared to the more reliable [O/H]_7773 (from O I 7771-5 triplet lines) or [O/H]_6300 (from [O I] 6300 line) obtained in this study. Comparing the updated [O/Fe] ratios with the theoretically predicted surface abundance changes caused by mixing of nuclear-processed products dredged-up from the interior, we concluded that the oxygen deficiency in these red giants is insignificantly marginal (only by <~0.1 dex), which does not contradict the expectation from the recent theoretical simulation. This consequence of reasonable consistency between theory and observation also applies to the extent of peculiarity in [C/Fe] and [Na/Fe], which were also examined by reanalyzing the previous equivalent-width data of C I 5052/5380 and Na I 6160 lines.