Low-Mass Companions Orbiting δ Scuti Stars

Abstract

A substantial fraction of exoplanets exist around M-type stars, but few systems have host stars hotter than the Sun and even fewer orbit hot stars that pulsate.

The discovery of planets around pulsating stars provides a laboratory that offers the opportunity for the most precise studies of exoplanet systems. To search for new systems, an exclusively UK-based exoplanet and variable star survey using UCLan’s Moses Holden Telescope (MHT) was established — MOSES: MHT Optical Star and Exoplanet Survey. The MOSES field was formed of a 4 × 4 tiled array covering a 7-square degree region of the sky observing over 65,000 stars, 60-s per exposure strategy in the Johnson V -band. Of all stars observed during the initial 10-month MOSES campaign, just 13 were flagged to be of a variable nature with a confidence of >5�. 10 of these 13 stars were sufficiently observed to be designated as MOSES Objects of Interest (MOIs), but the data were not of the quality required for the stellar and planetary modelling desired for the work in this thesis. As such, we ruled out a ground-based observing approach using this data set.

The Transiting Exoplanet Survey Satellite (TESS) data were scoured for δ Scuti host stars, suitable targets as their signature pulsation frequencies allow them to be well-modelled. Main-sequence � Scuti host stars are extremely rare: only seven are known in the literature. Two TOIs were chosen to be studied in detail, TIC 409934330 and TIC 156987351. A grid of stellar models was built with software mesa, based on a calculated range of stellar parameters; each model was then set pulsating by the software gyre. In parallel, the TESS data for each target were analysed in Period04 to extract, separately, the orbital frequency (and its harmonics) and the pulsation frequencies of the two targets. Finally, a chi-squared analysis was employed to determine the best-fit model star for each target, by comparing the respective gyre frequencies with the TESS-extracted frequencies. The final models produced a mass and radius for each target of 2.10 ± 0.19M� and 4.72 ± 0.44R�, and 1.63 ± 0.15M� and 3.05 ± 0.28R�, respectively, amongst other parameters.

Spectroscopic ground-based observations of these targets using the Southern African Large Telescope (SALT) provided spectra that were analysed in the software iSpec, determining a Teff of 8062±29K for TIC 409934330 and 7231±63K for TIC 156987351. The software exofastv2 used these parameters to deduce the nature of the transiting companions: TIC 409934330 hosts a 2.5+31.0 −2.1 MJ, 1.03+0.85 −0.49 RJ “hot-Jupiter”, and TIC 156987351 hosts a 0.39+7.90 −0.33MJ, 0.97+0.56 −0.67 RJ “hot-Saturn”.
Hence, we have added two more examples to the δ Scuti host star population.

Populating the HR diagram with all nine now-known examples, it became clear they all existed within a very narrow region of parameter space that we labelled “The δ Scuti Cone” — furthermore, all nine stars lay at the very base of the δ Scuti region in the classical instability strip. We explained this using the hypothesis that the more luminous δ Scuti stars disrupt their proto-planetary discs with their stellar winds before their planets can form. In addition, we included � Pictoris in our final investigation since it is a well-studied pre-main sequence δ Scuti host star, and its position on the HR diagram is consistent with our sample. We explored the theoretical δ Scuti instability strip according to Dupret et al. (2004), but show that it is too narrow and does not agree with our population, which agrees better with the empirical δ Scuti instability strip of Murphy et al. (2019).

It is recommended to investigate, in detail, the remaining planet candidates that lie in and around the The δ Scuti Cone, to ascertain their physical properties and environments that support, or otherwise, the Cone’s presence: such results could lead to defining a new sub-region on the classical instability strip.