Constraints and observational manifestations of failed solar eruptions in a toroidal magnetic cage

Context. Observations show that many solar eruptions remain confined in the magnetic configuration of strong overlying magnetic fields, which is commonly referred to as the magnetic cage. Aims. Confined eruptions under strong, poloidal overlying fields have been widely studied. In contrast, the confined eruption scenario under strong external toroidal fields remains unknown. Methods. We used three-dimensional magnetohydrodynamic simulations to systematically study confined eruptions in a toroidal magnetic cage, focusing on the roles of the Lorentz force and magnetic reconnection, as well as their observational manifestations, such as flare ribbons and loops. We further applied the test-particle method with the guiding-centre approximation to synthesize hard X-ray sources, enabling comparison between thermal and nonthermal responses. Results. Our results show that overlying toroidal magnetic fields are crucial in confining eruptions. They generate strong return currents that produce a significant downward Lorentz force, suppressing the rise of the flux rope. Simultaneously, they drive the large-angle rotation of the rope, triggering reconnection with the overlying fields and ultimately causing its breakup. The synthesized EUV images display multi-ribbon flare structures with highly sheared loops with a global “cowboy-hat-like” shape. Additionally, comparisons with hard X-ray sources reveal that thermal and nonthermal responses are not co-spatial, in which return current is a major accelerator to energetic electrons. Conclusions. The simulations clarify how the magnetic cage constrains solar eruptions. First, the downward Lorentz force related to return current effectively suppresses eruptions, explaining why confined flares tend to occur in electric-current neutralized active regions. Second, we demonstrate that toroidal-field-induced force is the key driver for rotation and confinement of the flux rope. This explains why many filament eruptions with rotation fail despite being torus-unstable. Finally, we suggest that the global morphology of flare loops (cowboy-hat–like or “saddle-like”) and the shearing degree of flare loops can serve as useful diagnostics to distinguish confined from eruptive flares.