Daniela Turis-Gallo (Universidad de Valparaiso, Chile)
Massive stars play a fundamental role in stellar evolution, feedback processes, and the chemical enrichment of galaxies. However, determining their fundamental parameters remains challenging because strong rotational effects, such as stellar deformation and gravity darkening, significantly modify their observed spectra. I will present a model-based spectroscopic analysis of rapidly rotating massive stars that consistently accounts for stellar oblateness and gravity darkening. Using synthetic spectra generated with the ZPEKTR code, we derive stellar parameters—in particular, the inclination angle and equatorial rotation velocity—from photospheric lines. I will analyse the intrinsic limitations of the ZPEKTR code as a function of stellar rotation rate, identifying the parameter space in which the models remain reliable and in which systematic uncertainties become significant. I will also show the application of the code to a sample of Be stars observed with high spectral resolution. Our results are compared with independent determinations from the literature, including interferometric measurements, providing a critical assessment of the accuracy of spectroscopic methods.
Juan Pablo Hidalgo (University of Rome)
Large-scale magnetic fields have been observed in about 10% of main-sequence early-type stars. Notably, chemically peculiar Ap/Bp stars can host surface magnetic fields with mean strengths between 200 G and 30 kG. Unlike late-type stars, whose magnetic fields have complex geometries and are likely generated by convective dynamos, the observed magnetic fields of early-type stars have simpler geometries, and are stable over long timescales, with virtually no variability over several decades. Because these stars have thick radiative envelopes and convective cores, surface dynamos are unlikely to account for the observed magnetism. Consequently, the origin of these magnetic fields remains uncertain. In this talk, I will review recent progress in some of the main theories proposed to explain the magnetism in early-type stars, exploring dynamos hosted by their convective cores, and their interaction with fossil fields inherited from earlier evolutionary stages. Furthermore, I will discuss ongoing simulations of pre-main-sequence stellar evolution aimed at constraining the structure and properties of these primordial fields.
Leda Berni (INAF- Osservatorio Astrofisico di Arcetri (Florence, Italy))
The Λ-CDM scenario predicts that the Milky Way assembled hierarchically, through the accretion of smaller systems such as globular clusters and dwarf galaxies. Although these systems can be disrupted by tidal interactions, their remnants survive as stellar streams that retain chemical and dynamical signatures of common origin. We developed CREEK, a machine-learning pipeline that learns similarity relations between stars using siamese neural networks, models their relational structure with graph neural networks autoencoders, and identifies substructures through density-based clustering (OPTICS). Applied to halo stars, our method recovers 80% of known globular clusters in the dataset, re-identifies established stellar streams, and reveals a candidate new substructure. We also detect two distinct populations within the Gaia Enceladus stream, possibly linked to multiple accretion passages through the Milky Way. The CREEK thus represents an objective and data-driven method for the selection of stars belonging to streams and to stellar structures in general, and the results highlight the capabilities of machine learning to find relations that would escape a classical search.
