A numerical model of cohesion in planetary rings

We present a numerical method that incorporates particle sticking in simulations using the N-body code pkdgrav to study motions in a local rotating frame, such as a patch of a planetary ring. Particles stick to form non-deformable but breakable aggregates that obey the (Eulerian) equations of rigid-body motion. Applications include local simulations of planetary ring dynamics and planet formation, which typically feature hundreds of thousands or more colliding bodies. Bonding and breaking thresholds are tunable parameters that can approximately mimic, for example, van der Waals forces or interlocking of surface frost layers. The bonding and breaking model does not incorporate a rigorous treatment of internal fracture; rather the method serves as motivation for first-order investigation of how semi-rigid bonding affects the evolution of particle assemblies in high-density environments. We apply the method to Saturn’s A ring, for which laboratory experiments suggest that interpenetration of thin, frost-coated surface layers may lead to weak cohesive bonding. These experiments show that frost-coated icy bodies can bond at the low impact speeds characteristic of the rings. Our investigation is further motivated by recent simulations that suggest a very low coefficient of restitution is needed to explain the amplitude of the azimuthal brightness asymmetry in Saturn’s A ring, and the hypothesis that fine structure in Saturn’s B ring may in part be caused by large-scale cohesion. This work presents the full implementation of our model in pkdgrav, as well as results from initial tests with a limited set of parameters explored. We find a combination of parameters that yields aggregate size distribution and maximum radius values in agreement with Voyager data for ring particles in Saturn’s outer A ring. We also find that the bonding and breaking parameters define two strength regimes in which fragmentation is dominated either by collisions or other stresses, such as tides. We conclude our study with a discussion of future applications of and refinements to our model.