Asteroseismic modelling of gravito-inertial modes in γ Doradus pulsators
Joey Mombarg (Institute of Astronomy, KU Leuven)
The data from the Kepler nominal mission have enabled the detection of gravito-inertial modes in sequences of consecutive radial order in about 600 γ Doradus (γ Dor) pulsators. The fact that these modes are sensitive to the deep interior of stars, makes γ Dor pulsators excellent candidates for improving the current incomplete prescriptions of the transport of angular momentum and chemical elements, both important ingredients in stellar evolution theory. To that extent, forward modelling of gravito-inertial modes to derive stellar masses, ages, and mixing efficiencies throughout the radiative envelope, helps us to understand the possible origins of these transport mechanisms in intermediate-mass main-sequence stars.
High-dimensional grid-based forward modelling quickly becomes computationally unfeasible for a large sample of stars due to the large number of required models. This can be circumvented by training neural networks on a set of stellar evolution and pulsation models to interpolate the observables in a high-dimensional parameter space. I will present a novel technique for modelling gravito-inertial modes in γ Dor pulsators, using deep learning, which relies on the individual observed mode periods, combined with spectroscopy, and Gaia luminosities as input. This method is combined with a framework which considers the correlated nature of the pulsations. The extracted (core) masses, ages, and core-overshooting parameters, combined with the rotation rates, allows for the investigation of any correlations present in the sample. Furthermore, I will discuss the effect of adding atomic diffusion (including radiative levitation) in the stellar models on the predicted pulsation frequencies, and compare the predicted surface abundances with observed ones.