Search for a command to run...
A bstract We consider the constraints from Supernova 1987A on particles with small couplings to the Standard Model. We discuss a model with a fermion coupled to a dark photon, with various mass relations in the dark sector; millicharged particles; dark-sector fermions with inelastic transitions; the hadronic QCD axion; and an axion-like particle that couples to Standard Model fermions with couplings proportional to their mass. In the fermion cases, we develop a new diagnostic for assessing when such a particle is trapped at large mixing angles. Our bounds for a fermion coupled to a dark photon constrain small couplings and masses ≲ 200 MeV, and do not decouple for low fermion masses. They exclude parameter space that is otherwise unconstrained by existing accelerator-based and direct-detection searches. In addition, our bounds are complementary to proposed laboratory searches for sub-GeV dark matter, and do not constrain several benchmark-model targets in parameter space for which the dark matter obtains the correct relic abundance from interactions with the Standard Model. For a millicharged particle, we exclude charges between 10 −9 –few×10 −6 in units of the electron charge, also for masses ≲ 200 MeV; this excludes parameter space to higher millicharges and masses than previous bounds. For the QCD axion and an axion-like particle, we apply several updated nuclear physics calculations and include the energy dependence of the optical depth to accurately account for energy loss at large couplings. These corrections allow us to rule out a hadronic axion of mass between 0.1 and a few hundred eV, or equivalently to put a bound on the scale of Peccei-Quinn symmetry breaking between a few×10 4 and 10 8 GeV, closing the hadronic axion window. For an axion-like particle, our bounds disfavor decay constants between a few×10 5 GeV up to a few×10 8 GeV, for a mass ≲ 200 MeV. In all cases, our bounds differ from previous work by more than an order of magnitude across the entire parameter space. We also provide estimated systematic errors due to the uncertainties of the progenitor.