Abstract

Contributed Talk - Splinter StarFormation   (MW-1250)

Stella: A flexible single stellar feedback model

Ulrich Steinwandel
Max Planck Institute for Astrophysics

We present Stella, a model for the self-consistent co-evolution of the multiphase interstellar medium (ISM) and a resolved stellar population in dwarf galaxies, implemented in the meshless finite-mass code GIZMO. Stella resolves star formation at the level of individual stars: each star particle is a single star whose mass is drawn from a fully-sampled Kroupa initial mass function and accreted from its natal gas within an accretion radius, eliminating the averaged stellar-population approximation. A star's mass, birth metallicity, and age set its luminosity, lifetime, nucleosynthetic yields, and compact-remnant type through pre-computed stellar-evolution tables. Feedback is treated channel-by-channel — photoionization, far-ultraviolet (Lyman–Werner and photoelectric) radiation propagated on the gravity tree, stellar winds, and individual core-collapse and Type Ia supernovae — coupled to a non-equilibrium hydrogen/CO chemical network with metal-line, molecular, and dust cooling. We track 27 elements with turbulent metal diffusion, an evolving dust-to-gas ratio, and cosmic-ray heating, so that the thermal and chemical state of the gas, the local radiation field, and the stellar population evolve together while conserving baryonic mass to machine precision. We demonstrate the model on isolated low-metallicity dwarfs, decomposing the contributions of each feedback channel to regulating star formation, shaping the thermal phase structure of the ISM, and driving outflows. Because feedback is resolved star-by-star, Stella naturally captures the intrinsic stochasticity of star formation in the low-mass regime, where sampling of the IMF and the clustering of massive stars dominate the dynamics.