The spectroscopic indistinguishability of red giant branch and red clump stars

Context. Stellar spectroscopy provides useful information on the physical properties of stars such as effective temperature, metallicity and surface gravity. However, those photospheric characteristics are often hampered by systematic uncertainties. The joint spectrosismo project (APOGEE+Kepler, aka APOKASC) of field red giants has revealed a puzzling off set between the surface gravities (log g) determined spectroscopically and those determined using asteroseismology, which is largely dependent on the stellar evolutionary status. Aims. Therefore, in this letter, we aim to shed light on the spectroscopic source of the off set. Methods. We used the APOKASC sample to analyse the dependencies of the log g discrepancy as a function of stellar mass and stellar evolutionary status. We discuss and study the impact of some neglected abundances on spectral analysis of red giants, such as He and carbon isotopic ratio. Results. We first show that, for stars at the bottom of the red giant branch where the first dredge-up had occurred, the discrepancy between spectroscopic log g and asteroseismic log g depends on stellar mass. This seems to indicate that the log g discrepancy is related to CN cycling. Among the CN-cycled elements, we demonstrate that the carbon isotopic ratio (C-12/C-13) has the largest impact on stellar spectrum. In parallel, we observe that this log g discrepancy shows a similar trend as the C-12/C-13 ratios as expected by stellar evolution theory. Although we did not detect a direct spectroscopic signature of C-13, other corroborating evidences suggest that the discrepancy in log g is tightly correlated to the production of C-13 in red giants. Moreover, by running the data-driven algorithm (the Cannon) on a synthetic grid trained on the APOGEE data, we try to evaluate more quantitatively the impact of various C-12/C-13 ratios. Conclusions. While we have demonstrated that C-13 indeed impacts all parameters, the size of the impact is smaller than the observed off set in log g. If further tests confirm that C-13 is not the main element responsible of the log g problem, the number of spectroscopic effects remaining to be investigated is now relatively limited (if any).