Simula'ng Photoevapora've Mass Loss from Hot Jupiters in 3D 

Anjali Tripathi1, atripathi@cfa.harvard.edu

Kaitlin Kratter2, Mark Krumholz3, Ruth Murray-Clay1

1 Harvard-Smithsonian Center for Astrophysics 2 JILA/University of Colorado, Boulder 3 University of California, Santa Cruz

References 

Ionization front

Stellar Flux

(not to scale)

Ionization front propagating through ambient gas

log ρ

 Goal To self-consistently model atmospheric escape in 3D, including photoionization heating.

•  UV spectra of hot Jupiters show significant absorption of stellar emission during transit.

•  This absorption is attributed to planet atmospheres that puff up and escape into space, due to heating by photoionizing stellar flux.

•  The geometry of heating, tidal locking, and orbital motion predict an asymmetric outflow structure.

•  Existing atmospheric escape models are either 1D or do not include photoionization heating.

Mo'va'on 

We use the ATHENA (Stone et al, 2008) code for hydrodynamics. The planet is at the center of a 3D Cartesian grid. Stellar radiation is input along one face of the box.

1. Computa'onal Setup 

• We use multiple levels of Static Mesh Refinement (SMR) to resolve the planet’s exosphere.

• The grid is divided and then distributed onto separate processors for parallel computation with MPI.

Hydrostatic atmosphere

Inner boundary (Hydro eq.s not solved here)

Ambient gas

Planet

log ρ

x

y

2D density profile

2. Planet We model the planet as a point mass potential with a dense neutral hydrogen atmosphere.

•  The atmosphere is isothermal, in hydrostatic equilibrium.

• Ambient gas has uniform density, pressure-matched to the outer edge of the atmosphere.

•  We do not solve the fluid equations in the planet’s interior, at radii within our inner boundary condition.

2D density slice

Schematic of our simulation box

2x resolution at each level of refinement

20 R p

20 Rp

Stellar Flux

(not to scale)

Static Mesh Refinement (SMR)

Decomposition for MPI parallelization

3. Stellar Radia'on Incident stellar radiation ionizes the gas through the radiative transfer algorithm of Krumholz et al. (2007), updated for use with SMR & MPI.

Neutral gas Ionized

gas

Numerical viscosity sets the simulated shock front thickness. With higher resolution from SMR, the front’s thickness decreases and the density increases.

SMR Region

Finer resolution in the SMR region allows better agreement between the simulated and analytic location of the ionization front. Error bars are given by the size of a grid cell.

• We model the stellar radiation as a planar source of monochromatic flux.

• Our algorithm includes photoionization heating, recombination cooling, and optically thin heating and cooling curves.

The planar ionizing radiative transfer algorithm, based on Krumholz et al. (2007), now features SMR and MPI capabilities for hydrodynamic simulations with ATHENA.

Consistency Check We verify that at late times, the location of the simulated ionization front matches analytic expectations for a D-type ionization front, when using SMR & MPI.

Time evolution of ionization front propagation

This material is based upon work supported by the National Science Foundation Graduate Research

Fellowship under Grant No. DGE-1144152.

4. Outlook 

• We plan to include the stellar wind and Coriolis force, to better characterize the asymmetric structure of the atmospheric outflow.

-Krumholz et al, 2007, ApJ, 671, 518 -Murray-Clay et al, 2009, ApJ, 693, 23 -Stone et al, 2008, ApJS, 178, 137 -Stone & Proga, 2009, ApJ, 694, 205 -Tremblin & Chiang, 2013, MNRAS, 428, 2565T

Our 3D hydrodynamics simulation with ionizing radiative transfer will allow us to model the launching and asymmetric structure of atmospheric escape.

Artist’s conception of a hot Jupiter undergoing mass loss. (ESA/NASA)

Existing models of mass loss Our simulation •  3D • Outflow launching by explicit

photoionization heating

•  (Stellar wind) •  (Coriolis force)

•  1D with explicit heating to launch the outflow

or

•  2D with stellar wind interaction

Ioni

zati

on

fro

nt lo

cati

on

[cm

]

SMR Region

Time [s]

Coarse resolution Fine resolution Analytic approximation

• Outflow confined by the stellar wind (Tremblin & Chiang, 2013)

•  Asymmetric column density predictions (Stone & Proga, 2009)

(Tremblin & Chiang, 2013)

•  Full 3D asymmetric structure •  Synthetic spectra