Escape of asteroids from the main belt

Abstract

Aims. We locate escape routes from the main asteroid belt, particularly into the near-Earth-object (NEO) region, and estimate the relative fluxes for different escape routes as a function of object size under the influence of the Yarkovsky semimajor-axis drift. Methods. We integrated the orbits of 78 355 known and 14 094 cloned main-belt objects and Cybele and Hilda asteroids (hereafter collectively called MBOs) for 100 Myr and recorded the characteristics of the escaping objects. The selected sample of MBOs with perihelion distance q > 1.3 au and semimajor axis a < 4.1 au is essentially complete, with an absolute magnitude limit ranging from HV < 15.9 in the inner belt (a < 2.5 au) to HV < 14.4 in the outer belt (2.5 au < a < 4.1 au). We modeled the semimajor-axis drift caused by the Yarkovsky force and assigned four different sizes (diameters of 0.1, 0.3, 1.0, and 3.0 km) and random spin obliquities (either 0 deg or 180 deg) for each test asteroid. Results. We find more than ten obvious escape routes from the asteroid belt to the NEO region, and they typically coincide with low-order mean-motion resonances with Jupiter and secular resonances. The locations of the escape routes are independent of the semimajor-axis drift rate and thus are also independent of the asteroid diameter. The locations of the escape routes are likewise unaffected when we added a model for Yarkovsky-O'Keefe-Radzievskii-Paddack (YORP) cycles coupled with secular evolution of the rotation pole as a result of the solar gravitational torque. A Yarkovsky-only model predicts a flux of asteroids entering the NEO region that is too high compared to the observationally constrained flux, and the discrepancy grows larger for smaller asteroids. A combined Yarkovsky and YORP model predicts a flux of small NEOs that is approximately a factor of 5 too low compared to an observationally constrained estimate. This suggests that the characteristic timescale of the YORP cycle is longer than our canonical YORP model predicts.