The intermediate neutron capture process. VI. Proton ingestion and i-process in rotating magnetic asymptotic giant branch stars

TL;DR

The paper examines how rotation and magnetic mixing influence i-process nucleosynthesis in PIEs within low-metallicity AGB stars, using STAREVOL with a large nuclear network and a Tayler-Spruit dynamo calibrated to asteroseismic core rotations. It shows that rotation without magnetic fields strongly suppresses the i-process via primary $^{14}$N production and subsequent $^{22}$Ne poisoning, whereas an asteroseismically calibrated magnetic dynamo couples the core and envelope, preventing excess $^{14}$N and restoring i-process yields to resemble non-rotating cases. Across [Fe/H] = $-2.5$ and $-1.7$ and initial masses of $1$ and $1.5\ M_{\odot}$, PIE-driven nucleosynthesis proceeds similarly in magnetic-rotating and non-rotating models, with detailed outcomes for heavy-element production, fluorine, and sodium that depend on timing of the PIE and prior $^{14}$N synthesis. The results highlight the important role of angular-momentum transport physics in AGB nucleosynthesis and suggest that additional transport mechanisms beyond the Tayler-Spruit dynamo may be required to fully capture observed spin properties and their impact on nucleosynthesis.

Abstract

The intermediate neutron-capture process (i-process) can occur during proton ingestion events (PIEs), which may take place in the early evolutionary phases of asymptotic giant branch (AGB) stars. We investigate the impact of rotational and magnetic mixing on i-process nucleosynthesis in low-metallicity, low-mass AGB stars. We computed AGB models with [Fe/H] = $-2.5$ and $-1.7$ and initial masses of 1 and 1.5 $M_{\odot}$ using the STAREVOL code, including a network of 1160 nuclei coupled to transport equations. Rotating models incorporate a calibrated Tayler-Spruit (TS) dynamo to account for core rotation rates inferred from asteroseismic observations of solar-metallicity sub-giants and giants. Initial rotation velocities of 0, 30, and 90 km s$^{-1}$ were considered, along with varying assumptions for magnetic mixing. We find that rotation without magnetic fields strongly suppresses the i-process due to the production of primary $^{14}$N, which is subsequently converted into $^{22}$Ne $-$ a potent neutron poison during the PIE. Including magnetic fields via the TS dynamo restores the models close to their non-rotating counterparts: strong core-envelope coupling suppresses shear mixing and prevents primary $^{14}$N synthesis, yielding i-process nucleosynthesis similar to non-rotating models. We also find that rotational mixing during the AGB phase is insufficient to affect the occurrence of PIEs. Proton ingestion event-driven nucleosynthesis proceeds similarly in asteroseismic-calibrated magnetic rotating AGB stars and non-rotating stars, producing identical abundance patterns.

Figures (7)

• Figure 1: Kippenhahn diagram illustrating key features of a PIE in a non-rotating 1 $M_{\odot}$, [Fe/H] $=-2.5$ AGB model, computed with the STAREVOL code. Grey regions indicate convective zones. The dashed green (blue) line marks the location of maximum H-burning (He-burning) energy. The colour map shows the neutron density. The time between the peak neutron density and the split is $\Delta t_1 \simeq 100$ hr, while the total duration of the sequence shown is $\Delta t_2 \simeq 0.1$ yr.
• Figure 2: Surface (dashed lines, $\Omega_{\rm s}$) and core (solid lines, $\Omega_{\rm c}$) rotation rates as a function of surface gravity for 1.1 $M_{\odot}$, solar-metallicity models with $v_{\rm ini} = 5$ km s$^{-1}$. Models are computed without (black) and with the Tayler instability, using $n=1$ and various values of the calibration constant $C_{\rm T}$ (green, red, and blue). The thick blue lines show the surface and core rotation of the 1.1 $M_{\odot}$, solar-metallicity model from eggenberger22 (E22) computed with GENEC ($n=1$, $C_{\rm T}=216$). Large black-filled (open) symbols indicate observed surface (core) rotation rates of sub-giant stars deheuvels14, while the smaller black dots show core rotation rates of RGB stars gehan18.
• Figure 3: Angular velocity profiles for rotating 1.0 $M_{\odot}$, [Fe/H] $=-2.5$ models for various values of $C_T$. The structure corresponds to the moment when the convective envelope reaches its deepest extent during the first dredge-up.
• Figure 4: Diffusion coefficients $D_{\rm shear}$ and $D_{\rm eff}$ between the convective thermal pulse and envelope, just prior to the PIE for 1 $M_{\odot}$, [Fe/H] $=-2.5$, $v_{\rm ini} = 30$ km s$^{-1}$ models with $C_T = 0$ (no TS dynamo), $C_T = 50$, and $C_T = 216$. The profiles of angular velocity ($\Omega$) and hydrogen mass fraction ($X_H$, scaled by $10^9$ and $10^3$, respectively) are also shown. Grey- and orange-shaded areas indicate convective and overshoot zones, respectively.
• Figure 5: Mass fractions of heavy elements in the convective thermal pulse, just before (dotted lines) and after (solid lines) the main proton ingestion for our 1 $M_{\odot}$, [Fe/H] $=-2.5$ models. Left panel: Non-rotating model (black) and models with $v_{\rm ini} = 30$ km s$^{-1}$ and $C_T = 0$ (magenta), 1 (green), 50 (red), and 216 (blue). Right panel: Non-rotating model (black) and models with $v_{\rm ini} = 90$ km s$^{-1}$ and $C_T = 50$ (red) and 216 (blue).