Weakened Inspirals I: High Mass Ratio Common Envelope Interactions in RGB Stars

TL;DR

This work demonstrates that high mass ratios ($q$) can weaken common envelope inspirals in RGB binaries, producing wider post-interaction separations and enabling fallback-driven circumbinary discs. Using 3D SPH simulations with varying $q$, resolution, and equation of state, the authors show that larger $q$ lengthens the pre-inspiral mass-transfer phase and yields final separations up to $\sim40\,R_\odot$ (with extreme $q$ values approaching larger radii), while disc formation is more plausibly achieved through bound fallback material at radii of $0.5$–$5$ au rather than through $L_2$/$L_3$ ejection. The fallback discs have masses around $10^{-2}\,M_\odot$ and short viscous timescales, aligning with observed circumbinary discs around post-RGB/AGB systems, though the widest observed binaries remain challenging to reproduce. Energy and angular momentum are conserved to tight tolerances, underscoring the robustness of the simulations and clarifying the role of recombination energy in envelope unbinding. Overall, these results offer a viable pathway toward weakened CE outcomes at high $q$, while highlighting remaining gaps and the need for further studies on AGB donors and additional physics.

Abstract

The common envelope (CE) interaction between an expanding giant star and a compact companion typically leads to a rapid orbital decay, ending in either a merger or the formation of a close binary. However, the existence of post-red giant and post-asymptotic giant branch binaries with separations of 100 to 800 Rsun challenges this standard picture, as these systems appear to have experienced strong interactions without undergoing a classic CE inspiral. In this work, we investigate the effect of high mass ratio, q = M2/M1, on the CE inspiral using three-dimensional hydrodynamical simulations performed with the smoothed particle hydrodynamics code PHANTOM. The primary is a 0.88 Msun, 90 Rsun red giant branch star, while the companion masses span q = 0.68 to 1.5. Higher mass ratios lead to wider post-CE separations, with a maximum of approximately 40 Rsun. The pre-CE mass transfer phase is longer for larger companion masses, and for q greater than or equal to 1 the inspiral becomes significantly more stable, broadly consistent with analytical expectations. This phase is not fully converged with respect to numerical resolution, and higher resolution simulations are expected to further increase its duration and stability. Although higher q systems show enhanced mass loss through the L2 and L3 Lagrange points, we find that circumbinary discs are more likely to form from fallback of bound envelope material. Fallback times are short, of order a few hundred years, and fallback radii lie well outside the binary, between 0.5 and 5 au, where discs are expected to spread efficiently through viscous torques. While high mass ratio systems produce wider post-interaction separations, these remain smaller than those observed. In contrast, fallback-formed discs have properties consistent with observed circumbinary discs.

Figures (14)

• Figure 1: Illustration summarising the three evolutionary phases of post-RGB and post-AGB binaries. The interaction stage refers to a type of 'weak' CE interaction, or a grazing envelope evolution interaction Socker2015. After the envelope ejection and the end of the central binary's inspiral, the second phase is defined by the formation of a stable circumbinary disc. The third phase is much slower and is the one we observe today. The disc is stable and the orbit has been cleared, even if accretion onto the central binary continues.
• Figure 2: Cross sections of density in the orbital plane of the 68H (left), 85H (centre), and 100H (right) simulations. Each column is a time sequence starting with two moments before the inspiral (top two rows), and ending with the start ($t_i$) and end ($t_f$) of the inspiral (bottom two rows). Each box is approximately 7 au in size.
• Figure 3: Top panel: binary core separation as a function of time for the twelve simulations (see Table \ref{['tab:CE_summary']}). The circles and triangles are the start and end of the inspiral, respectively, as determined by the criterion $|\frac{\dot{a}}{a}| \geq \frac{1}{15} \textrm{max}|\frac{\dot{a}}{a}|$Reichardt2019. Note that this criterion is not adopted for the $q=1.5$ simulations, either due to a very shallow inspiral (150L and 150MH), or the lack of inspiral (150H). Due to this shallow inspiral, we have opted for an alternative criterion (see text). Extrapolating from the time taken for the 100L and 100H simulations to inspiral, the computational cost for continuing the 150H simulation is currently unfeasible. Bottom panel: the evolution of the bound mass for each simulation. Circles and triangles have the same meaning as in the upper panel, while the stars denote the time at which the resolution-dependent mass unbinding is estimated to start.
• Figure 4: Distribution of bound mass ($K+U<0$) throughout simulations 68H (top left), 85H (top right), 100H (bottom left), and 150H (bottom right). The pixels are binned at approximately 10 days in width, and 5 R$_{\odot}$ in height, where we calculate the average energy of the gas within that radial bin, at that time step. Top panel: normalised orbital separation (blue) and the bound envelope (red). The vertical lines spanning the two plots denote, from left to right, the start (solid) and end (dashed) of the inspiral. These lines correspond respectively to the circle and triangle in Figure \ref{['fig:evo_vs_time']}.
• Figure 5: As in Figure \ref{['fig:LH_bound_inspiral']}, but for the 68MH, the 85MH, the 100MH and the 150MH simulations.