Abstract

The dominant uncertainty in late-stage binary evolution is not nuclear physics, but the coupling between mass loss and orbital dynamics. As low- and intermediate-mass stars evolve along the giant branches, they develop extended, pulsation-supported atmospheres and dust-driven winds with substantial mass-loss rates. In a binary system, these outflows are generically non-isotropic: the companion gravitationally focuses the wind, drives spiral shocks, enables enhanced accretion, and mediates angular momentum exchange through both direct torques and circumbinary reservoirs. The commonly adopted assumption of isotropic, Jeans-mode mass loss — widely adopted in population synthesis and secular evolution codes — becomes highly questionable in this regime.

In this talk, I will present a combined observational and 3D radiation–hydrodynamic–chemical modelling approach to quantify the angular momentum budget of evolved binaries. Spatially resolved ALMA data constrain the density and velocity field of the outflow, while Hipparcos and Gaia astrometry constrain orbital architectures. These empirical constraints are coupled to self- consistent 3D simulations that resolve wind launching, gravitational deflection, dust formation, and companion accretion. By measuring the specific angular momentum carried by the escaping material relative to the orbital angular momentum, we can directly assess how mass-loss–driven torques modify semi-major axis and eccentricity evolution. This provides a physically grounded pathway from resolved flow dynamics to secular orbital evolution, and challenges the validity of classical isotropic wind prescriptions in predicting the fate of interacting giant binaries.

Because evolved binaries are progenitors of white dwarf pairs, chemically peculiar stars, and interacting compact systems, their physics propagates directly into population synthesis, supernova rates, and galactic chemical enrichment. Moreover, the formation of circumbinary reservoirs links this interaction regime to disk dynamics and potentially to second-generation planet formation. Understanding how winds couple to orbits is therefore not a specialised problem in stellar evolution, but a key ingredient in tracing how binaries shape stellar populations and galaxies across cosmic time.