Engulfment of Eccentric Planets by Giant Stars: Hydrodynamics and Light Curves

TL;DR

The paper investigates how an eccentric giant planet engulfed by a red-giant envelope can produce luminous transients resembling LRNe. Using 3D hydrodynamic simulations with Athena++ and post-processing to generate light curves, it shows that repeated planet–star encounters generate shocks that eject envelope gas, boosting luminosity by orders of magnitude and creating a hydrogen-recombination plateau followed by dust-driven dimming. The results yield ejecta masses of a few ×10^{-3} to a few ×10^{-3}–6×10^{-3} M_⊙ and reveal viewing-angle dependent brightness with quasi-periodic peak structures tied to the spiral shocks, while the opacity treatment (MESA vs analytic) significantly affects the plateau temperature and dust timing. Limitations include the absence of radiation transport and a simplified equation of state, yet the findings have implications for interpreting LRNe-like transients and events such as ZTF SLRN-2020, and motivate further work on dust physics and magnetohydrodynamic effects in planetary engulfment scenarios.

Abstract

Recent observations suggest that planetary engulfment by a giant star may produce radiation that resembles subluminous red novae. We present three-dimensional hydrodynamical simulations of the interaction between an eccentric $5 \,M_J$ giant planet and its $1\,M_\odot$ red-giant host star. The planet's pericenter is initially $60\%$ of the stellar radius and is fully engulfed after tens of orbits. Once inside the stellar envelope, the planet generates pressure disturbances that steepen into shocks, ejecting material from the envelope. We use post-processing to calculate the light curves produced by planetary engulfment. We find that the hot stellar ejecta enhances the stellar luminosity by several orders of magnitude. A prolonged hydrogen recombination plateau appears when the ejecta cools to about $10^4\,\rm{K}$. The late-time rapid dimming of the light curve follows dust formation, which obscures the radiation. For planets with lower eccentricity, the orbital decay proceeds more slowly, although the observable properties remain similar.

Figures (8)

• Figure 1: Opacities of the analytical model (solid lines) and from the MESA tables (dashed lines) as a function of temperature for selected densities (as labeled, in $\mathrm{g/cm^3}$).
• Figure 2: Slices of the density (top row) and temperature (bottom row) through the orbital plane for the fiducial simulation. The left two columns correspond to the first planet–star encounter, followed by snapshots during the planet’s spiral-in and after its disruption. The stellar core is marked by a large white circle, while the planet is indicated by a white (black) circle in the upper (lower) panels. The grid scale corresponds to $1\,R_*$, as shown by the white scale bar in the first panel. The time unit is $t_\mathrm{dyn}$, given by Eq. \ref{['eq:tdyn']}.
• Figure 3: Temporal evolution of the spherically averaged stellar density, pressure, temperature, and sound speed profiles. In the bottom right panel, the spherically averaged gas velocity is shown alongside the sound speed. Different times are indicated by color.
• Figure 4: Time evolution of planet's orbital radius $r_p$ (light blue line, top panels), radial luminosity $L_\mathrm{rad}$ (red line, top panels), spherically averaged photospheric radius (light green line, middle panels) and temperature (yellow line, middle panels), observed luminosity (bottom panels) from three different viewing angles ($L_{x}$, dark blue line; $L_y$, brown line; $L_z$, dark green line). The left panels show the observable properties obtained from the analytical formula of opacity, while the right panels are the corresponding results using the MESA opacity tables. In the top panels, the initial star radius and inner boundary radius are indicated by the black dotted and dashed lines, respectively. The radial and observed luminosities are normalized to the initial star luminosity $L_*$. After the planet is disrupted when it reaches the inner boundary at $\sim 175\, t_\mathrm{dyn}$, the orbital radius is set to the radius of the inner boundary ($0.3\, R_*$).
• Figure 5: The orbital decay rate, $|\dot{r}_{\mathrm{apo}}|$, as a function of the planetary radius $R_p$ (which is the softening radius). The blue circles show the results from Athena++ simulations differing only in the softening length. The stars mark the predictions from the semi-analytical model based on three choices of $L_\mathrm{int}$: orange for $\sqrt{R_*^2-r^2}$ (Ostriker), green for $H_\rho$ (O'Conner), and red for $H_{\rho,\,\hat{v}}$ (this work). The corresponding drag-force coefficient, $C_g$, is shown on the right axis. The fiducial case is indicated by an overlaid black cross.