Probing the Axion-Nucleon Coupling with Supergiant Stars

TL;DR

This work targets the axion-nucleon coupling by exploiting the 14.4 keV nuclear transition of ${}^{57}$Fe in the hot cores of supergiant stars. It develops a forward-modeling, Bayesian framework to search for a monochromatic 14.4 keV line from axion-to-photon conversion in the Galactic magnetic field, using NuSTAR observations of Betelgeuse and a detailed MESA-based stellar model to predict the axion flux. The analysis yields the strongest direct bound to date on the product $|g_{a\gamma} g_{aN}^{\mathrm{eff}}|$ for $m_a \lesssim 10^{-10}$ eV, with $|g_{a\gamma} g_{aN}^{\mathrm{eff}}| \le (1.2-2.7)\times 10^{-20}$ GeV$^{-1}$, and finds no significant line. The results demonstrate the viability of supergiants as axion laboratories and point to future gains from additional targets (e.g., M82) and improved Galactic magnetic field modeling, offering a complementary avenue to solar and terrestrial axion searches.

Abstract

A finite axion-nucleon coupling enables the production of axions in stellar environments via the thermal excitation and subsequent de-excitation of the $^{57}$Fe isotope. Given its low-lying excited state at 14.4 keV, $^{57}$Fe can be efficiently excited in the hot cores of supergiant stars, possibly leading to axions emission. The conversion of these axions into photons in the Galactic magnetic field results in a characteristic 14.4 keV line, potentially detectable by hard X-ray telescopes such as NASA's Nuclear Spectroscopic Telescope Array (NuSTAR). In this work, we present the first constraints on axion-nucleon couplings derived from \textsc{NuSTAR} observations of Betelgeuse and discuss the potential insights that could be gained from detecting this line in other nearby supergiants. Our results establish significantly more stringent bounds than those obtained from solar observations, setting a limit of $|g_{aγ} g_{aN}^{\mathrm{eff}}| < (1.2 - 2.7) \times 10^{-20}$ GeV$^{-1}$ for $m_a \lesssim 10^{-10}$ eV.

Figures (9)

• Figure 1: Exclusion limits from the Bayesian unbinned likelihood analysis at 95% confidence level (CL), in the case of massless axions compared to previous $^{57}$Fe studies. Our conservative results ($B_T = 1.4\ \mu G$) are displayed in dark red and a more optimistic bound, using $B_T = 3.0\ \mu G$, in a lighter red. Shown are the latest bounds on the axion-photon coupling $|g_{a\gamma}|$ from CAST CAST:2024eil (dashed) and results from the Nuclear Spectroscopic Telescope Array (NuSTAR) observations of Betelgeuse Xiao:2020pra and the Sun Ruz:2024gkl in gray, the solar cooling bound on the effective axion-nucleon coupling $|g_{aN}^{\mathrm{eff}}|$DiLuzio:2021qct in darker gray and results of searches for the $14.4$ keV line of $^{57}$Fe similar to this work from CAST CAST:2009jdc (dashed) and CUORE Li:2015tyq in a lighter gray.
• Figure 2: Expected axion signal in the two modules FPMA (left) and FPMB (right). Shown is the signal for an effectively massless axion ($m_a \lesssim 10^{-10}$ eV) assuming the maximally allowed value for the combination of the couplings (at 95% CL) $|g_{a\gamma}g_{aN}^{\mathrm{eff}}| = 2.7\times 10^{-20}$ GeV$^{-1}$. The spectra are binned in 1 keV increments and shown in the 9–21 keV range for visualization purposes, while the analysis was conducted on the full energy range using unbinned data.
• Figure 3: Exclusion limits from the Bayesian unbinned likelihood analysis at 95% CL for massive axions. Shown are bounds from searches for the $14.4$ keV line of $^{57}$Fe by CAST CAST:2009jdc and CUORE Li:2015tyq.
• Figure 4: Central abundances of hydrogen, helium, carbon, and oxygen for the supergiant star as a function of time. Tracking the abundances allows one to identify the key stages of the stellar evolution, like the onset of helium and carbon burning in the stellar core.
• Figure 5: Left: stellar core temperature as a function of time to core collapse for our simulation. Right: radial temperature profile of our benchmark $20M_\odot$ stellar model, taken at a time when the carbon burning phase is about to finish.