Small-scale collisions with big-scale effects: Direct numerical simulations of crystal interactions in dense suspensions and ramifications for magmatic differentiation

Numerous problems in the Earth sciences involve the dynamic interaction between solid bodies and viscous flow. The goal of this contribution is to develop and validate a computational methodology for modeling complex solid-fluid interactions with minimal simplifying assumptions. The approach we develop is general enough to be applicable in a wide range of geophysical systems ranging from crystal-bearing lava flows to sediment-rich rivers and aerosol transport. Our algorithm relies on a two-step projection scheme: In the first step, we solve the multiple-phase Navier-Stokes or Stokes equation, respectively, in both domains. In the second step, we project the velocity field in the solid domain onto a rigid-body motion by enforcing that the deformation tensor in the respective domain is zero. An important component of the numerical scheme is the accurate treatment of collisions between an arbitrary number of suspended solid bodies based on the impact Stokes number and the elasticity parameters of the solid phase. We perform several benchmark computations to validate our computations including wake formation behind fixed and mobile cylinders and cuboids, the settling speed of particles, and laboratory experiments of collision modes. Finally, we apply our method to investigate the competing effect of entrainment and fractionation in crystalline suspensions - an important question in the context of magma differentiation processes in magma chambers and magma oceans. We find that the properties and volume fraction of the crystalline phase play an important role for evaluating differentiation efficiency.