Title:On the seismic modelling of subgiant stars: testing different grid interpolation methods

Abstract:Context. The emergence of mixed modes during the subgiant phase, whose frequencies are characterized by a fast evolution with age, can potentially enable a precise determination of stellar properties, a key goal for future missions like PLATO. However, current modelling techniques often consider grids that lack the resolution to properly account for the fast mode frequency evolution, consequently requiring the use of interpolation algorithms to cover the parameter space in between the grid models when applying model-data comparison methods. Aims. We aim at reproducing the $\ell$=1 mode frequencies within the accuracy limits associated with the typical observational errors ($\sim$0.1 $\mu$Hz) through interpolation on a grid of subgiant models. Methods. With that aim, we used variations of a two-step interpolation algorithm, which considered linear and cubic splines interpolation methods and different age proxies (physical age, scaled age, and central density). Results. The best results were obtained using an algorithm that considers cubic splines interpolation along tracks, linear interpolation across tracks, and central density $\rho_\text{c}$ as the age proxy. This combination yielded, on average, an absolute error of 0.14 $\mu$Hz, but reached maximum absolute errors on the interpolated frequencies of 1.2 $\mu$Hz for some models, which is an order of magnitude higher than the typical observational errors. Furthermore, we investigated the impact on the accuracy of the interpolation from changes in the physical properties of the stars, showing, in particular, how the addition of core overshoot can affect significantly the interpolation results.