2023ApJ...943...95S

In this work, for the first time, we use seismic data as well as surface abundances to model the supergiant α-Ori, with the goal of setting an upper bound on the axion-photon coupling constant g_aγ. We find that, in general, stellar models with g_aγ ∈ [0.002; 2.0] × 10^–10 GeV^–1 agree with the observational data, but beyond that upper limit, we do not find stellar models that are compatible with the observational constraints and the current literature. From g_aγ = 3.5 × 10^–10 GeV^–1 on, the algorithm did not find any fitting models. Even so, all the axionic models considered present distinct internal profiles from the reference case, without axions. Moreover, as the axion energy losses become more significant, the behavior of the stellar models becomes more diversified, even with very similar input parameters. Nonetheless, the consecutive increments of g_aγ still show systematic tendencies, resulting from the axion energy losses. Moreover, we establish three important conclusions: (1) the increased luminosity and higher neutrino production are measurable effects, possibly associated with axion energy losses; (2) stellar models with axion energy loss show a quite distinct internal structure; and (3) the importance of future asteroseismic missions in observing low-degree nonradial modes in massive stars is emphasized-as internal gravity waves probe the near-core regions, where axion effects are most intense. Thus, more seismic data will allow us to constrain g_aγ better and to prove or dismiss the existence of axion energy loss inside massive stars.