WASP-12, shrouded in mystery or just cold gas?

TL;DR

This work tackles the puzzle of strong Mg II absorption toward WASP‑12 by constructing a complete line-of-sight density profile that combines a planet‑wind–driven torus, the stellar wind, the WASP‑12 astrosphere, and the interstellar medium (ISM). Using PLUTO-based hydrodynamics, plus a 3D ISM/ionization framework, the authors find the torus becomes dense and cool but remains predominantly Mg III, leaving Mg II largely supplied by the ISM along the 413 pc sightline. The total Mg II column is far short of the observed value of $N_{MgII}\approx2\times10^{17}$ cm$^{-2}$, with the torus contributing only a tiny fraction and the ISM commanding ~99.99% of the Mg II budget. They propose that a cooling instability in the torus could fragment gas into cold clumps with higher Mg II fractions, potentially reconciling the observations in future work, though this cooling was not included in the present model.

Abstract

Observations of the planet-hosting star WASP-12 show a distinctive depression in the \ion{Mg}{ii} and \ion{Ca}{ii} resonance lines. This has been interpreted as a marker of atmospheric loss from the close-in hot Jupiter WASP-12b and the resulting formation of a gas torus around the star. In this paper we quantify the \ion{Mg}{ii} absorption from this torus, compared to that provided by the stellar wind, the stellar astrosphere and the ISM. To do this we piece together the full density profile of \ion{Mg}{ii} from WASP-12 to an observer on Earth using a combination of hydrodynamical simulations and observations. We find that the bulk of the gas along the line of sight is contained within a dense torus close to WASP-12. However, the temperatures in this torus are sufficient to promote Mg into a doubly (\ion{Mg}{iii}) or higher ionized state. As a result, the singly ionized fraction (\ion{Mg}{ii}) is low. We find that most of the \ion{Mg}{ii} is not in the torus but in the ISM. Despite this, the total column density of \ion{Mg}{ii} is two orders of magnitude lower than required to explain observations of the system. To resolve this discrepancy, we note that the torus gas is at a temperature where it will cool efficiently. We speculate that the onset of the cooling instability will cause the torus to fragment, forming cold clumps with a higher fraction of \ion{Mg}{ii}, capable of explaining the observed absorption.

Figures (9)

• Figure 1: Volume rendering of the 3D gas density distribution (blue colour map) around WASP-12 after 60 orbits. The star is located in the centre (yellow), with the planet embedded in the torus, offset along the $x$-axis from the star and is highlighted by the black circle.
• Figure 2: Slice plots time-averaged over the last 10 orbits of the planet for three different views: top down (left), side on in the plane of the planet's position (middle) and side on perpendicular to the plane of the planet's position. Time averaging is done in the rotating frame of the planet, which is held fixed at orbital separation for the plots shown here. For this reason the downstream spiral structure behind the planet remains visible despite the time averaging. From top to bottom: density, temperature, velocity magnitude and tracer concentration.
• Figure 3: Top: time series of (hydrogen) gas mass-build-up in the simulation domain. At late times the curve has flattened off, indicating convergence of the simulation. Bottom: block average (over 64 outputs at a time) of the rate of change of total mass in the simulation. This method smooths out the stochastic fluctuations in the total mass, ensuring $\mathrm{d}m/\mathrm{d}t > 0$ at all times. We use this as an indicator of the convergence of the simulation. While the value of $\mathrm{d}m/\mathrm{d}t$ has not reached zero by $t=75 t_{\rm orbit}$, we deem that the rate of mass-build-up has slowed sufficiently to consider the simulation converged.
• Figure 4: Radial mass density (top) and temperature (bottom) profile at times $t=0$ (dotted, the initial, pure Parker wind from the star) and $t=75 \rm \ t_{\rm orbit}$ (dot-dashed) through the orbital plane of the planet. Also shown are profiles averaged over the last 10 orbits, both along the LOS (solid) and through the plane of the planet (dashed). The location of the planet is marked by the vertical line. Profiles shown here are computed at an azimuthal angle of 180 degrees from the planet.
• Figure 5: Number density and temperature of the astrosphere of WASP-12. The system is located at the origin. The simulation is conducted in the co-moving frame with the system, as such the ISM inflows from the right had side of the plot at a velocity of 23.72 km s$^{-1}$ and density of $2.212\times10^{-1} \ \mathrm{cm^{-3}}$ ($2.22 \times 10^{-25} \ \mathrm{g} \ \mathrm{cm^{-3}}$) (see text in Section \ref{['sec:astrosphere_sim']} for details of these values). This inflow collides with the spherically expanding stellar wind resulting in the formation of a termination shock, and astropause. The structure is assumed to be cyclindrically symmetric about the x-axis. This allows us to conduct the simulation in 2D. The LOS to the observer is annotated by the white line.