Simulations of two-planet systems through all phases of stellar evolution: implications for the instability boundary and white dwarf pollution

Abstract

Exoplanets have been observed at many stages of their host star's life, including the main sequence (MS), subgiant and red giant branch stages. Also, polluted white dwarfs (WDs) likely represent dynamically active systems at late times. Here, we perform 3-body simulations which include realistic post-MS stellar mass loss and span the entire lifetime of exosystems with two massive planets, from the endpoint of formation to several Gyr into the WD phase of the host star. We find that both MS and WD systems experience ejections and star-planet collisions (Lagrange instability) even if the planet-planet separation well-exceeds the analytical orbit-crossing (Hill instability) boundary. Consequently, MS-stable planets do not need to be closely-packed to experience instability during the WD phase. This instability may pollute the WD directly through collisions, or, more likely, indirectly through increased scattering of smaller bodies such as asteroids or comets. Our simulations show that this instability occurs predominately between tens of Myr to a few Gyrs of WD cooling.A planet's life may be split into four distinct stages: 1) formation and concurrent dynamical excitation, 2) main sequence (MS) evolution, 3) evolution during post-MS stellar phase changes, and 4) white dwarf (WD) evolution. The first stage generally lasts no longer than 0.1% of the entire MS lifetime. The second stage is relatively dynamically quiescent, with only occasional but often important scattering interactions. In the third stage, the planet is subject to dynamical changes due to the star's violent actions as it becomes a giant. In the final stage, the star has become a WD, and the planet again enters and remains in a phase of relative dynamical quiescence occasionally punctuated by scattering interactions or external forcing. This general picture, which does not include possibilities such as the capture of free-floating planets, planetary destruction due to supernovae, or multiple host stars, describes the life cycle of the vast majority of known exoplanets.The volume of planetary literature investigating the ⋆

An extreme-AO search for giant planets around a white dwarf

“…GD 50 is the most distant white dwarf in the figure, and yet our limit on GD 50 is still the most stringent. (Veras et al 2013), so any planets within 6.5 au were possibly destroyed. Therefore, it is unlikely that there are any giant planets more massive than 2.8 M J around GD 50 now.…”

Remnant planetary systems around bright white dwarfs

“…An easier way to detect remnant planetary systems around WDs is to look for the tidally disrupted remains of exoplanets, moons, and asteroids in the form of circumstellar debris discs (Debes & Sigurdsson 2002;Jura 2003;Kilic et al 2006;Farihi, Jura & Zuckerman 2009;Veras et al 2013Veras et al , 2014aVeras, Jacobson & Gänsicke 2014b). Consider a planetary system consisting of a WD, an asteroid-belt analogue and a Jupiter analogue.…”

“…Rafikov & Garmilla (2012) find that the dust sublimation temperature may be increased due to the vapour pressure of sublimated dust at the disc's inner rim. Poynting-Robertson drag carries dust from the tidal radius inward towards the WD, where viscous torques cause the sublimated dust to accrete on to the WD's surface (Rafikov 2011a,b;Metzger, Rafikov & Bochkarev 2012;Veras et al 2013). This accretion results in a spectroscopically detectable pollution of the otherwise pristine WD atmosphere.…”

The fate of exomoons in white dwarf planetary systems

“…Only a few studies have integrated suites of self-consistent multiplanet simulations across the main sequence, giant branch and white dwarf phases. Veras et al (2013) and Mustill, Veras & Villaver (2014) performed two-planet and three-planet simulations, respectively. However, for computational reasons, both studies modelled stars with main-sequence masses of 3 M or greater.…”