Abstract DGP2026-98

Planets and stars form from the same proto-stellar material. Hence the stellar refractory elemental abundances are assumed to be strongly linked to rocky planet interiors. This is also found for the refractory elemental abundances of the Sun and Earth. For exoplanets, this compositional link has been recently suggested and explored in small demographic studies. However, the sample of rocky planets around metal-poor stars is limited. This makes it challenging to find and validate potentially vital chemical trends. We present a novel machine learning approach to identify planet hosting stars of interest to fill in the lack of well-characterised small planets around metal-poor stars. This algorithm leverages stellar abundances from the large stellar surveys of APOGEE and GALAH to classify host stars chemically followed by a search for planets around these golden targets. 

We present newly characterised systems containing Ultra-Short Period Super-Earths and Sub-Neptunes around compositionally-diverse main sequence stars. These planetary systems were observed with transit photometry and radial velocity allowing us to measure their radius and masses precisely. We modelled their interior structures for which we further developed tools to link stellar abundances to the core and mantle mass fractions of the planet. These give insights into the elemental abundance ratios of the planets.  Hence, we directly study the connection between the planet and host star abundances. 

These new discoveries significantly add to the sample of rocky planets around thick disk stars and will be followed by further detections and characterisations from our machine learning algorithm. Placing our new and upcoming systems in demographic context, we are now able to further explore the link between stellar refractory elemental abundances and their impact on the interiors of small planets in a larger and more diverse sample. This enables insights into rocky planet formation across the galaxy.